[68Ga]NS3-RGD and [68Ga] Oxo-DO3A-RGD for imaging αvβ3 integrin expression: synthesis, evaluation, and comparison

Nuclear Medicine and Biology 40 (2013) 65–72

Contents lists available at SciVerse ScienceDirect

Nuclear Medicine and Biology journal homepage: www.elsevier.com/locate/nucmedbio

[ 68Ga]NS3-RGD and [ 68Ga] Oxo-DO3A-RGD for imaging αvβ3 integrin expression: synthesis, evaluation, and comparison☆ Peter A. Knetsch a, Milos Petrik a, b, Christine Rangger a, Gesine Seidel c, Hans-Jürgen Pietzsch c, Irene Virgolini a, Clemens Decristoforo a, Roland Haubner a,⁎ a b c

Department of Nuclear Medicine, Innsbruck Medical University, Innsbruck, Austria Institute of Molecular and Translational Medicine, Palacky University, Olomouc, Czech Republic Helmholtz-Zentrum Dresden Rossendorf, Institute of Radiopharmacy, Dresden, Germany

a r t i c l e

i n f o

Article history: Received 22 June 2012 Revised in revised form 15 September 2012 Accepted 20 September 2012 Keywords: 68 Ga RGD αvβ3 Molecular imaging Angiogenesis

a b s t r a c t Introduction: 68Ga-labeled RGD peptides in combination with PET allow non-invasive determination of αvβ3 integrin expression which is highly increased during tumor-induced angiogenesis. The aim of this study was to synthesize and evaluate two RGD peptides containing alternative chelating systems, namely [68Ga]NS3RGD and [ 68Ga]Oxo-DO3A-RGD and to compare their in vitro and in vivo properties with [68Ga]DOTA- and [68Ga]NODAGA-RGD. Methods: Syntheses of both radiotracers followed standard SPPS protocols. For in vitro characterization distribution coefficients, protein binding abilities, serum stabilities, and αvβ3 integrin binding affinities were determined. For in vitro tests as well as for the biodistribution assay αvβ3 positive human melanoma M21 and αvβ3 negative M21-L cells were used. Results: 68Ga-labeling of NS3-RGD resulted in good radiochemical purity, whereas HPLC analysis showed two peaks with a ratio of 1:6 for [ 68Ga]Oxo-DO3A-RGD. Distribution coefficients were −3.4 for [ 68Ga]Oxo-DO3ARGD and −2.9 for [68Ga]NS3-RGD. Both radiotracers were stable in PBS solution at 37 °C for 2 h but lack stability in human serum. Protein binding was approximately 40% of the total activity for [ 68Ga]NS3-RGD and 70% for [68Ga]Oxo-DO3A-RGD, respectively, resulting in high blood pool activities. Biodistribution assays confirmed these findings and showed an additional high uptake in liver and kidneys, especially for [68Ga]NS3RGD. Furthermore, [68Ga]Oxo-DO3A-RGD showed nearly the same activity concentrations in αvβ3 positive and αvβ3 negative tumors. Conclusions: [68Ga]Oxo-DO3A-RGD and [ 68Ga]NS3-RGD have inferior characteristics compared to already existing 68Ga-labeled RGD peptides and thus, both are not suited to image αvβ3 integrin expression. Of all our tested RGD peptides, [68Ga]NODAGA-RGD still possesses the most favorable imaging properties. Moreover this study shows that the use of appropriate chelators to achieve good targeting properties of 68Ga-labeled biomolecules and careful in vitro and in vivo evaluation including comparative studies of different strategies are essential components in designing an effective imaging agent for PET. © 2013 Elsevier Inc. All rights reserved.

1. Introduction An approach to treat cancer is based on the inhibition of tumorinduced angiogenesis. One of the key players involved in this angiogenic process is the αvβ3 integrin. This integrin is overexpressed on activated endothelial cells during blood vessel formation where it affects tumor growth, local invasiveness, and the metastatic potential [1]. Pathological angiogenesis is not only found in cancer [2], but also many other diseases are characterized by modified

☆ Conflict of interest: The authors declare that they have no conflict of interest. ⁎ Corresponding author. Universitätsklinik für Nuklearmedizin, Medizinische Universität Innsbruck, Anichstr. 35, A-6020 Innsbruck, Austria. Tel.: +43 512 504 80069; fax: +43 512 504 6780069. E-mail address: [email protected] (R. Haubner). 0969-8051/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.nucmedbio.2012.09.006

angiogenesis, e.g. psoriasis [3], atherosclerosis [4], restenosis [5], and rheumatoid arthritis [6]. It has been demonstrated in preclinical as well as in clinical studies that radiolabeled RGD peptides and positron emission tomography (PET) allow monitoring of αvβ3 expression which holds a great potential in drug discovery and new therapeutic approaches [7]. Techniques allowing non-invasive imaging to select patients and to control therapeutic success are of great interest. One of the most intensely evaluated compounds so far is [ 18F]Galacto-RGD [8] which specifically binds to the αvβ3 integrin and shows very good pharmacokinetics. Highly favorable biodistribution characteristics were found for [ 18F]Galacto-RGD in patient studies also demonstrating that the standard uptake value (SUV) correlates with αvβ3 expression [9–11]. Nevertheless, the synthesis of this compound is complex and time consuming (overall synthesis time is more than

66

P.A. Knetsch et al. / Nuclear Medicine and Biology 40 (2013) 65–72

two hours), making an automated process almost impossible. With this in mind the use in clinical routine is limited [12]. Due to the increasing availability of 68Ge/ 68Ga generators, 68Galabeled compounds, especially peptides, are coming more and more into the focus of interest. Advantages are the access of a PET isotope without the need of an in-house cyclotron and the straightforward labeling strategies which are optimal for synthesis automation. Thus, we started to develop 68Ga-labeled RGD-peptides to overcome the production problems 18F-labeled derivatives have to deal with. Due to the positive experience with [ 68Ga]DOTA-TOC, the initial selection of an appropriate chelating systems was focused on 1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid (DOTA) leading to the development of [ 68Ga]DOTA-RGD as an alternative to [ 18F]GalactoRGD [13]. This tracer showed receptor specific uptake in tumor tissue but also high blood protein bound activity resulting in inferior imaging properties. To overcome this problem we started to synthesize and evaluate a series of RGD-peptides with different chelating systems. This includes 1-Oxa-4,7,10-triazacyclododecane-5S-(4-isothiocyanatobenzyl)-4,7,10-triacetic acid (Oxo-DO3A), a 1,4,7triazacyclononane-N,N′,N″-triacetic acid (NOTA) derivative [14], and 3-Benzylsulfanyl-2-[bis(2-benzylsulfanylethyl)amino]propionic acid (NS3-COOH) [15]. Out of this series data on [ 68Ga]NODAGA-RGD, which includes a NOTA derivative offering six coordination sites for Ga 3+ complexation, have already been presented [16]. This study showed, that this tracer can be produced even at room temperature, showing high affinity to the αvβ3 integrin, possesses a high metabolic stability, and has a low serum protein binding. Here we present the syntheses and evaluation of the remaining peptides [ 68Ga]Oxo-DO3A-RGD and [ 68Ga]NS3-RGD of this series and compared the in vitro and in vivo results of these new tracers with [ 68Ga]DOTA-RGD and [ 68Ga]NODAGA-RGD. 2. Materials and methods 2.1. General All purchased substances and solvents were of reagent grade and were used without further purification. 9-Fluorenylmethoxycarbonyl (Fmoc)-protected amino acids were obtained from Novabiochem (La Jolla, CA, USA). Trityl chloride polystyrene (TCP) resin was purchased 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 Gen-Script 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. The p-SCN-Bn-Oxo-DO3A chelator was a gift of Macrocyclics Inc (Dallas, TX, USA). For electrospray ionization mass spectrometry (ESI-MS) analysis 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) was used. Reversed-phase high-performance liquid chromatography (RPHPLC) analysis was carried out by using two different pump systems — (1) a Dionex P680 or (2) a Dionex P580 HPLC pump with an UVD 170 U UV/VIS detector (Dionex, Germering, Germany) and a Bioscan radiometric detector (Bioscan, Washington D.C., USA). An ACE Nucleosil C18 3 μm, 150×3.0 mm column (Bartelt GmbH, Innsbruck, Austria), flow rate 0.5 mL/min, and a Nucleosil 120-5C18, 250×4.0 mm column (SRD, Vienna, Austria), flow rate 1.0 mL/min, respectively, and UV detection at 220 nm were employed with the following acetonitrile (ACN)/water — 0.1% trifluoroacetic acid (TFA)

gradients: for (1) 0–1.5 min 0% ACN, 1.5–18.0 min 0–50% ACN, 18.021.0 50–100% ACN (gradient A) and for (2) 0–2.0 min 10% ACN, 2.017.0 min 10–50% ACN (gradient B). Isolation of the peptide via semipreparative RP-HPLC was performed using a Gilson 322 HPLC pump with 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 10% ACN, 1.0– 25.0 min 10%–80% ACN (gradient C). Radioactivity of the samples was measured using a 2480 Automatic Gamma Counter Wizard 2 3″ (PerkinElmer, Vienna, Austria). The 68Ga generator was purchased from Eckert & Ziegler (Berlin, Germany) with a nominal activity of 1100 MBq and was eluted with 0.1 N HCl (Sigma-Aldrich Handels GmbH; Vienna, Austria) using the fractionated elution approach [17]. 2.2. 3-Benzylsulfanyl-2-[bis(2-benzylsulfanylethyl)amino]propionic acid 2,5-oxopyrrolidin-1-yl ester (N(Sbz)3COONHS) The precursor 3-benzylsulfanyl-2-[bis-(2-benzylsulfanylethyl) amino]propionic acid (N(Sbz)3COOH) was prepared from S-benzylcysteine by reaction with 1-benzylsulfanyl-2-bromoethane and subsequent ester hydrolysis. Synthesis details can be found in reference [16]. 3-Benzylsulfanyl-2-[bis-(2-benzylsulfanylethyl) amino]propionic acid (200 mg, 390 μmol) was dissolved in 3 mL of N,N-dimethylformamide (DMF), and a solution of 80 mg (390 μmol) N,N’-dicyclohexylcarbodiimide in 2 mL of DMF was added at 0 °C. After the mixture was stirred for 20 min, 45 mg (390 μmol) of N-hydroxy succinimide (NHS), dissolved in 1 mL of DMF, was added. The mixture was stirred for 4 h at 0 °C and overnight at room temperature (RT). The formed precipitate was separated by filtration, the supernatant was evaporated, and the remaining oil was purified by column chromatography (silica gel 60, n-hexane/ethyl acetate 1/1) to give the product as white oil (yield: 210 mg, 88%). ESI-MS (m/z) calculated (found): 610 (610) for [M+H] + 2.3. Cyclo(–Arg(Pbf)-Gly-Asp(OtBu)-DPhe-Lys–) The linear RGD-peptide was synthesized on a solid support using a TCP resin and Fmoc protocols. Protecting groups were 2,2,4,6,7pentamethyldihydrobenzofuran-5-sulfonyl (Pbf) for arginine, tertbutyl (tBu) for aspartic acid, and benzyloxycarbonyl (Z) for lysine. Cyclization of the peptide was performed in N,N-dimethyl formamide (DMF) in the presence of diphenylphosphoryl azide (DPPA) and sodium hydrogen carbonate (NaHCO3). Subsequent deprotection of lysine was carried out under hydrogen atmosphere in the presence of an activated charcoal palladium catalyst in N,N-dimethylacetamide (DMA) solution resulting in cyclo(–Arg(Pbf)-Gly-Asp(OtBu)-DPhe-Lys–). Synthesis details can be found in reference [18]. 2.4. Deprotection of cyclo(–Arg(Pbf)-Gly-Asp(OtBu)-DPhe-Lys–) Complete deprotection of 100 mg cyclo(–Arg(Pbf)-Gly-Asp(OtBu)was carried out by dissolving the peptide in 5 mL solution consisting of TFA/H2O/triisopropylsilane (95: 2.5: 2.5). The reaction mixture was allowed to react for 10 h at RT. Hereafter, the solvent was evaporated, the residue diluted in a minimum amount of DMF, and crude cyclo(–Arg-Gly-Asp-DPhe-Lys–) (c(RGDfK)) was precipitated by addition of diethyl ether. DPhe-Lys–)

2.5. Conjugation of (N(Sbz)3COONHS) and final deprotection of the chelator To a mixture of 10 mg (14 μmol)) c(RGDfK) and 10 μL triethyl amine, dissolved in 1 mL DMF, 16 mg (27 μmol) N(Sbz)3COONHS was

P.A. Knetsch et al. / Nuclear Medicine and Biology 40 (2013) 65–72

added to the solution. The mixture was allowed to react at ambient temperature for 4 h resulting in cyclo(–Arg-Gly-Asp- D PheLys(NHCON(Sbz)3–) (N(Sbz)3-RGD). After removal of the solvent the crude product was dissolved in a mixture of 500 μL acetonitrile and 500 μL water. Purification was carried out by semipreparative HPLC: Knaur Wellchrom HPLC pump, stationary phase 90 A, 4 μm, 250×10 mm (Phenomenex Jupiter Proteo), mobile phase: A: ACN 0,1% TFA, B: H2O 0,1% TFA, gradient from 5 to 65% A within 40 min, flow rate: 4 mL/min, 300 μL injected. tR =33 min N(Sbz)3-RGD, yield: 60%. A solution of 10 mg (9μmol) N(Sbz)3-RGD in 3 mL of dry THF was dropped to about 5 mL of liquid ammonia. Sodium was added in small portions until the solution remained deep blue for 15 min. The solvents were removed in an argon stream. The white residue was dissolved in a mixture of 1 mL methanol, 1.5 mL water and 500 μL acetonitrile. The solution was acidified with 500 μL 2 M HCl (pH ~1) to give the crude NS3RGD. Purification was carried out by semipreparative HPLC: Knaur Wellchrom HPLC pump, stationary phase 90A, 4 μm, 250×10 mm (Phenomenex Jupiter Proteo), mobile phase: A: ACN 0,1% TFA, B: H2O 0,1% TFA, gradient from 5 to 65% A within 40min, flow rate: 4 mL/min, 300 μL injected. tR = 18 min NS3-RGD, yield: 60%. ESI-MS: m/z [M+H]+ =844.3 [C34H54N10O8S3; exact mass 843 (calculated)]. 2.6. Conjugation of Oxo-DO3A To 80 mg of c(RGDfK) (0.13 mmol) dissolved in 1 mL DMF, 120 mg Oxo-DO3A (0.20 mmol) was added to the solution. The mixture was allowed to react for 20 h resulting in cyclo(–Arg-Gly-Asp-DPheLys(Oxo-DO3A)–) (Oxo-DO3A-RGD), which was purified via RPHPLC preparation (gradient C): tR =9.6 min, ESI-MS: m/z [M+H] + = 1098.6 [C49H71N13O14S; exact mass 1097 (calculated)]. 2.7. [ 68Ga]Oxo-DO3A-RGD Labeling was carried out using the fractionated elution method. To 20 μL of peptide solution (1 μg/μL in water), 1 mL of activity (50– 150 MBq 68Ga eluate in 0.1 M HCl), and 100 μL sodium acetate solution (1.9 M) were added. The labeling mixture was allowed to react for 10 min at 95 °C. For in vivo as well as in vitro studies the solution was used without further purification. RP-HPLC: tR =19.7 min (gradient A). 2.8. [ 68Ga]NS3-RGD For radiolabeling 40 μg of peptide was dissolved in 240 μL of aqueous Tris(2-carboxyethyl)phosphine hydrochloride (TCEP·HCl) solution (5 mM). After 30 min of vigorous shaking at RT, the newly formed gel film was separated via centrifugation at 100 rcf for 3 min, the supernatant was taken off, and the residue washed with ethanol (30 μL) (two times). Alcoholic washing solutions and the supernatant were combined, mixed with 400 μL phosphate buffer pH 7.0 (0.067 M), and 1 mL of the main fraction of the 68Ga eluate (50– 150 MBq, in 0.1 M HCl) was added. Hereafter, the solution was adjusted to pH 7 using aqueous sodium hydroxide (1.0 M) and was allowed to react for 15 min at RT. [ 68Ga]NS3-RGD was used for in vitro and in vivo studies without further purification. RP-HPLC: tR = 10.5 min (gradient B). 2.9. [ 68Ga]DOTA-RGD and [ 68Ga]NODAGA-RGD Radiolabeling of these tracers was carried out as previously described [13,16]. 2.10. Distribution coefficient (logD) [ 68Ga]Oxo-DO3A-RGD and [ 68Ga]NS3-RGD (approximately 10 kBq in 50 μL, 1 nmol peptide) were diluted in 450 μL PBS, respectively. 500 μL of octanol were added to the peptide solution, the mixtures

67

vortexed for 15 min at 1400 rpm and centrifuged for 2 min at 2000 rcf. Subsequently, aliquots of the aqueous and the octanol layer were collected, measured in the gamma counter, and logD values were calculated (n=5). 2.11. Stability studies Stability studies were carried out by incubating either [ 68Ga]OxoDO3A-RGD or [ 68Ga]NS3-RGD in 2 mL of PBS (1.5 MBq) or 2 mL (5 MBq) of fresh human serum. The radiolabeled compounds were incubated for 30, 60, and 90 min at 37 °C. At the selected points, an aliquot of the PBS solution was analyzed via HPLC, whereas the serum aliquot was fixed on a sep-pak column, washed with 500 μL water and eluted with ACN containing 0.1% TFA. The solvent was removed in vacuo and the residue dissolved in 500 μL PBS. HPLC analysis gradient A was used and extraction efficacy was determined by dividing the activity of the sep-pak eluate by the total activity used in this assay. 2.12. Protein binding Protein binding abilities for either [ 68Ga]Oxo-DO3A-RGD or [ 68Ga] NS3-RGD were evaluated by incubation of the radiolabeled compounds for several time points (30, 60, and 120 min) at 37 °C in fresh human serum. Subsequently, the solutions were passed through a size exclusion spin column (MicroSpin™ G-50 columns, GE healthcare, Buckinghamshire, UK). Protein binding abilities were determined by measuring the activity bound to the column (non protein bound) and the activity in the eluate (protein bound) in a gamma counter. 2.13. Isolated receptor binding affinity (IC50 values) In vitro binding affinities of cyclo(–Arg-Gly-Asp-DTyr-Val–) (c(RGDyV)), Oxo-DO3A-RGD, and NS3-RGD were determined by using immobilized integrin αvβ3 (Millipore-Chemicon, Temecula, CA, USA) and 125I-echistatin (Amersham-Pharmacia Biotech, Vienna, Austria) as radioligand. The detailed procedure was reported previously [19]. Briefly, 96-well plates (Nunc, Thermo Fisher Scientific, Vienna, Austria) were coated for 16 h at 4 °C with the αvβ3 integrin. Hereafter, the immobilized receptors were incubated with 125Iechistatin and increasing concentrations of the corresponding peptide (0.01–100 nM). The unbound fraction of radioligand was washed out and receptor-bound activity was obtained by treating the wells with sodium hydroxide solution (2 M). IC50 values were determined by fitting the per cent inhibition using Origin software (Northhampton, MA, USA). Three independent measurements were made. 2.14. Biodistribution 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). For the induction of tumor xenografts, M21 and M21-L cells were subcutaneously injected at a concentration of 5×10 6 cells/mouse and allowed to grow until tumors reached a size of 0.3–0.6 cm 3. To determine the αvβ3 receptor specific uptake Balb/c nu/nu mice (Charles River, Sulzfeld, Germany) bearing the human melanoma M21 in the right flank and αvβ3 negative M21-L (as a negative control) in the left flank were used (n=5 for each peptide). Mice were injected either with [ 68Ga] Oxo-DO3A-RGD (~1 MBq/animal, ~0.4 μg peptide) or [ 68Ga]NS3-RGD (~1 MBq/animal, ~0.5 μg peptide) intravenously in the tail vein. The animals were sacrificed by cervical dislocation 60 min post injection. Organs (heart, stomach, lung, spleen, liver, pancreas, kidneys, and intestine), blood, muscle tissue and tumors were removed and weighed. Activity uptake of the samples was measured in the gamma counter. Results were expressed as percentage of injected dose per gram tissue (%ID/g).

68

P.A. Knetsch et al. / Nuclear Medicine and Biology 40 (2013) 65–72

3. Results 3.1. Peptide synthesis and conjugation of the chelator Cyclo(–Arg(Pbf)-Gly-Asp(OtBu)-DPhe-Lys–) could be obtained in good yields. Synthesis of the RGD peptide was realized using solid phase peptide synthesis (SPPS) with Fmoc protocols. Cyclization was carried out under high dilution conditions, and deprotection of the lysine side chain under hydrogen atmosphere. Amidation of OxoDO3A with cyclo(–Arg-Gly-Asp-DPhe-Lys–) (c(RGDfK)) was carried out by in situ activation resulting in Oxo-DO3A-RGD (chemical purity of 95%). NS3-RGD was obtained by conjugation of cyclo(–Arg(Pbf)Gly-Asp(OtBu)-DPhe-Lys–) with N(Sbz)3COONHS via its activated carboxylic function (N-hydroxysuccinimidyl ester) and final deprotection using sodium in liquid ammonia. The achieved chemical purity was 97% as determined by HPLC. For the peptide synthesis scheme see Fig. 1. 3.2. Radiolabeling HPLC analysis of [ 68Ga]Oxo-DO3A-RGD (proposed structure see Fig. 2) revealed two peaks in a ratio of 1:6 which were not separated

for further studies (Fig. 3). [ 68Ga]NS3-RGD (proposed structure see Fig. 2) resulted in approximately 72% radiochemical yield (RCY) and more than 97% RCP after carrying out a sep-pak purification step. Pretreating the NS3-RGD peptide with the reducing agent, TCEP, enhanced the labeling yields to approx. 97% and consequently making a sep-pak purification unnecessary (Fig. 3). Specific activities ranged from 1 to 3 MBq/nmol for [ 68Ga]NS3-RGD and from 3 to 9 MBq/nmol for [ 68Ga]Oxo-DO3A-RGD. 3.3. In vitro characterization Distribution coefficients (logD) for [ 68Ga]Oxo-DO3A-RGD and [ Ga]NS3-RGD were −3.4 and −2.9, respectively, revealing a hydrophilic character of the peptides. Both investigated tracers were stable in PBS solution for the whole monitoring period of 1.5 h. However, for [ 68Ga]Oxo-DO3A-RGD 75% and 20% of intact peptide were found after 60 and 90 min incubation in fresh human serum, respectively (extraction efficacy of approx. 25%). [ 68Ga]NS3-RGD revealed an even lower stability in serum, just 50% of the intact tracer was found after 30 min, after 60 min no intact radiolabeled compound could be identified (extraction efficacy of approx. 59%). Comparison of stability assay data found for DOTA-RGD [13] and NODAGA-RGD [16] is shown in Fig. 4. The protein bound fraction for [ 68Ga]NS3-RGD was approx. 40% of total activity. Especially [ 68Ga]Oxo-DO3A-RGD had a very high protein binding value of approx. 70% of total activity for the whole monitoring period (Fig. 5). Increasing amounts of the peptides Oxo-DO3A-RGD, NS3-RGD, or c(RGDyV) successfully suppressed the binding of 125I-echistatin to the immobilized αvβ3 integrin receptor. IC50 values were 4.1±0.8 nM for Oxo-DO3A-RGD, 4.8±1.8 nM for NS3-RGD, and 1.3±0.5 nM for c(RGDyV) (reference), respectively. 68

3.4. In vivo characterization Biodistribution and specific tumor uptake were determined 60 min after injection of [ 68Ga]Oxo-DO3A-RGD and [ 68Ga]NS3-RGD, respectively, in nude mice bearing M21 as well as M21-L tumors. A comparison with the data found for DOTA-RGD [13] and NODAGARGD [16] is shown in Fig. 6. The in vivo studies revealed clear differences in the distribution pattern. [ 68Ga]NS3-RGD showed high affinity for αvβ3 integrin resulting in an absolute uptake in tumor tissue of 3.4±1.2 %ID/g and 1.2±0.7 %ID/g in the negative control. In contrast, [ 68Ga]Oxo-DO3ARGD showed nearly same activity accumulations in αvβ3 positive (5.9±1.0 %ID/g) as well as αvβ3 negative (6.4±1.1 %ID/g) tumors. High activity concentrations in liver (69.6±21.9 %ID/g) and kidneys (24.5±8.9 %ID/g) were found for NS3-RGD reflecting the results of the stability studies. [ 68Ga]NS3-RGD (3.3±1.7 %ID/g) revealed a three times lower accumulation in blood than [ 68Ga]OxoDO3A-RGD (11.3±3.8 %ID/g). In general, with exception of the activity concentration in blood, the activity distribution found for [ 68Ga]Oxo-DO3A-RGD ranged between approx. 2.7 and 5.2 % ID/g for all organs studied and were thus, much more homogenous as found for [ 68Ga]NS3-RGD (ranges form approx. 0.8–69.6 % ID/g).

4. Discussion [ 68Ga]DOTA-RGD [13] was the first compound out of a series of Ga labeled RGD peptides which were synthesized and evaluated by our group. This radiotracer is fast and easy to label, moreover radiolabeling can be carried out in a remote-controlled system. Additionally, [ 68Ga]DOTA-RGD showed high affinity for the αvβ3 integrin, but also high radioactivity associated with blood proteins was found. Biodistribution studies confirmed these findings showing 68

Fig. 1. (A) Coupling of the first amino acid; (B) Assembly of the linear peptide on the solid support; (C) Cleavage from the resin; (D) Cyclization; (E) Selective Removal of the Z-group; (F) Complete deprotection of cyclo(–Arg(Pbf)-Gly-Asp(OtBu)-DPhe-Lys–); (G) Conjugation of N(Sbz)3COONHS; (H) Final deprotection of the chelator resulted in NS3-RGD; (I) Conjugation of Oxo-DO3A resulted in Oxo-DO3A-RGD.

P.A. Knetsch et al. / Nuclear Medicine and Biology 40 (2013) 65–72

69

Fig. 2. Proposed structures of (A) [68Ga]NS3- and (B) [68Ga] Oxo-DO3A-RGD.

higher blood activity levels resulting in a lower blood/tumor ratio and consequently leading to inferior imaging properties found in dynamic microPET analysis compared to [ 18F]Galacto-RGD [13]. However, the reason for the high protein binding of [ 68Ga]DOTA-RGD is not quite clear. A potential explanation is that the Ga 3+ ion does not ideally fit in the DOTA cage due to its ion radius which seems not as suitable as other radiometals like 111In (62 pm for Ga 3+ versus 80 pm of In 3+) [20]. Due to the fact that gallium ions normally form 4 to 6 coordinated complexes, one carboxylic group of the DOTA chelator remains free which may interact with blood proteins (in comparison In 3+ ions form complexes with 8 coordinations) [20]. We found that [ 68Ga]DOTA-RGD showed a higher uptake in blood and also in the tumor than [ 111In]DOTA-RGD. In contrast, Dijkgraaf et al. [21] most recently demonstrated that 68Ga- and 111In-labeled DOTA-E-c(RGDfK) show nearly equal protein-bound activity values (b5%) as well as tumor uptake. However, due to the different analysis modalities and murine tumor model used a direct comparison is difficult. Anyway, based on our observations regarding DOTA-RGD it was of interest to investigate RGD-peptides conjugated to other chelating systems, including [ 68Ga]NODGA-RGD [16], [ 68Ga]NS3-RG, and [ 68Ga]Oxo-DO3A-RGD. Govindaswamy and co-workers [22] described the use of a tripodal tetradentate ligand NS3 (tris(2-mercaptobenzyl)amine which forms stable complexes with Fe 2+ and Fe 3+. This chelator can also be used for the complexion of Ga 3+ due to its chemical similarities with the Fe 3+ ion. The gallium ion is a slightly weaker hard (Lewis) acid, has a

smaller radius (62 pm for Ga 3+ versus 65 pm of Fe 3+) [20], and is a bit softer than the iron ion which is supposed to result in an even higher affinity for the softer thiolate moieties. Welch and co-workers [23] investigated tris(2-mercaptobenzyl) amine with In 3+ and Ga 3+ in solution and in the solid state. Protonation and metal binding constants were determined to help interpret the different in vivo behavior of the complexes as potential

Fig. 3. Radiochromatograms of (A) [68Ga]NS3-RGD and (B) [68Ga]Oxo-DO3A-RGD. [68Ga]NS3-RGD was labeled in good RCY and RCP, whereas HPLC analysis showed two peaks with a ratio of 1:6 for [68Ga]Oxo-DO3A-RGD which have not been separated for further studies.

70

P.A. Knetsch et al. / Nuclear Medicine and Biology 40 (2013) 65–72

Fig. 4. Comparison of serum stability of [68Ga]NS3-, [68Ga]Oxo-DO3A-, [68Ga]DOTA-, and [68Ga]NODAGA-RGD at 37 °C for 30, 60, and 90 min. Whereas [68Ga]DOTA- and [68Ga] NODAGA-RGD showed almost no sign of degradation for the whole monitoring period, [68Ga]NS3-and [68Ga]Oxo-DO3A-RGD lacked stability in serum. * % of intact tracer.

imaging probes. Interestingly, the Ga(III) complex of NS3 crosses the blood–brain-barrier and also accumulates in the heart. Obviously, as a small, neutral and lipophilic species, the complex meets the basic requirements for a potential brain imaging agent. The 68Ga complex was easily formed with a radiochemical purity of N95%. In vitro stability studies of the [ 68Ga]NS3 complex, determined in rat serum incubated at 37 °C, showed that more than 95% intact compound was found at 2 h by silica gel and reversed-phase radio-TLC [24]. However, in this form the chelator is not suitable for conjugation to targeting vectors like peptides and proteins. In a subsequent step the synthesis of bis(2-thio)benzyl-(2-thio-4aminobenzyl)amine, a bifunctional derivative of the NS3 core was introduced by Luyt and Katzenellenbogen [25]. According to the authors, the new NS3 derivative should enable the conjugation to peptides and allow the preparation of 68Ga-labeled radiopharmaceuticals. In a 13 step synthesis procedure phenylalanine was conjugated. However, 68Ga radiochemistry was not described. The complicated functionalization of the NS3 ligand, as well as the high lipophilicity due to its aromatic character, leading to a high accumulation in liver, makes this chelator obstructive for the purpose of peptide or protein conjugation. Here we present as a feasible alternative the tripodal ligand NS3COOH, a bifunctional chelator, which is predominantly used as a tetradentate ligand forming in combination with a monodentate coligand, such as isocyanide or phosphine, ‘4+1’ technetium(III) and rhenium(III) complexes, respectively [15]. The non-coordinating carboxyl group allows easy conjugation of targeting domains such as peptides or can serve as anchor for pharmacologically modifying groups [26]. Complexation of Ga 3+ by NS3-COOH results in a similar ‘4+1’ coordination geometry with water as monodentate co-ligand [27]. Additionally, like the different NOTA derivatives, NS3-COOH allows 68Ga-labeling at room temperature in high radiochemical yields. This finding, together with its easy synthesis, made the chelator interesting for comparative evaluation. One working hypothesis was, that the high protein binding tendency of [ 68Ga]DOTA-RGD could be due to the free carboxylate function of the complex. Thus, we have been looking for derivatives forming 68Ga-complexes with high thermodynamic stability avoiding such effects. Oxo-DO3A shows an almost similar stability constant for the Ga(III)-complex (log Km =21.3 [28]) compared to the [Ga]DOTAcomplex (log Km =21.33 [29]) but contains, in contrast to DOTA, only three carboxylic-functions, which are all involved in the complexation of the radiometal. These properties have convinced us to evaluate this chelating system as an additional alternative to the DOTA system. Ferreira et al. [30] described that the optimized reaction conditions for 68Ga-labeling of p-NO2-Bn-Oxo, a comparable Oxo-DO3A-system,

can be carried out within 5 min at room temperature with an RCY of 98.5%. These findings are contrary to our results. We obtained [ 68Ga] Oxo-DO3A-RGD by heating for 10 min at 95 °C and almost no labeling with 68Ga could be observed at RT. HPLC analysis of the labeled product resulted in two peaks in a 1 to 6 ratio. Due to the fact that the uncoordinated precursor was obtained in high chemical purity (single peak), an explanation of this peak splitting could be based on possible facial–meridial isomerism of the octahedral coordination expected for the Ga-DO3A system (N3O3-coordination). Comparison of these new compounds with DOTA-RGD [13] and NODAGA-RGD [16] showed that all tracers have comparable distribution coefficients and revealed, in in vitro binding studies, high affinity for the αvβ3 integrin. In contrast, [ 68Ga]NS3-RGD and [ 68Ga]Oxo-DO3A-RGD showed poor serum stability and clear differences in the biodistribution pattern. Only [ 68Ga]DOTA-RGD, [ 68Ga]NODAGA-RGD, and [ 68Ga]NS3-RGD showed a receptor specific uptake in the αvβ3 positive tumor. Of all the tested compounds [ 68Ga]NS3-RGD had the highest uptake in tumor tissue, which was approximately fourfold higher than that of [ 68Ga]NODAGA-RGD. A fourfold higher activity accumulation in the M21 tumor than in the

Fig. 5. Comparison of blood protein binding properties. Amount of protein-bound activity, found by size exclusion chromatography, after several time points of incubation of the corresponding tracers in human serum showed high accumulation of activity for [68Ga]NS3-RGD (approx. 40%) and even higher amounts for [68Ga]OxoDO3A-RGD (approx. 70%). Both are significantly higher than the values found for [68Ga] DOTA-RGD [13] and [68Ga]NODAGA-RGD [16].

P.A. Knetsch et al. / Nuclear Medicine and Biology 40 (2013) 65–72

71

where different chelating systems, including DOTA, NODAGA, OxoDO3A, and NS3-COOH are conjugated with c(RGDfK), indicates, a) that Oxo-DO3A and NS3-COOH are not well suited for complexing 68Ga, b) that the performance of DOTA may depend on the targeting vector used, and, that c) overall, NOTA derived chelators are the superior systems for complexing 68Ga. Moreover, out of this series [ 68Ga]NODAGA-RGD is the most promising compound for non-invasive monitoring of αvβ3 integrin expression and is a potential successor of [ 18F]Galacto-RGD not only for hospitals with no in-house cyclotron. Overall, this study shows the importance of using appropriate chelators to achieve good targeting properties of 68Ga-labeled biomolecules and, that careful in vitro and in vivo evaluation including comparative studies of different strategies is an essential component in designing effective imaging biomolecules for PET.

Acknowledgments

Fig. 6. Comparison of biodistribution data shows clear differences in the distribution pattern. Nude mice bearing the αvβ3-positive human melanoma M21 on the right flank and the negative control tumor M21-L on the left flank were used for biodistribution studies. Data were collected 60 min after injection and showed for all tracers, except [68Ga]Oxo-DO3A-RGD, specific accumulation in the tumor.

reference M21-L tumor confirms the specific uptake of [ 68Ga]NS3RGD, whereas [ 68Ga]Oxo-DO3A-RGD showed similar accumulations in αvβ3 positive as well as αvβ3 negative tumors. The low metabolic stability of the NS3-chelating systems may be the main reason for the high uptake in liver and kidneys. Increasing blood pool activities correlate with the corresponding high protein bound activity found for both tracers. Whereas the Ga-DOTA and the Ga-Oxo-DO3A complex show similar stability constants, [ 68Ga]DOTA-RGD and [ 68Ga]Oxo-DO3A-RGD differ completely regarding their in vitro and in vivo behavior. In particular, Ferreira et al. [30] studied transmetallation of their Oxo-DO3A-system and found that at 37 °C almost all of the 68Ga was transferred to apo-transferrin within 15 min. This result indicates that the high protein bound fraction found for [ 68Ga] Oxo-DO3A-RGD could also be due to transmetallation to apotransferrin, one of the major blood glycoproteins. Altogether, the high activity accumulation of [ 68Ga]NS3-RGD in liver and kidneys and for [ 68Ga]Oxo-DO3A-RGD an additionally high activity concentration in blood in accordance with the data found in the stability assays indicate that these tracers are not well suited for imaging of αvβ3 expression. Other groups used Oxo-DO3A for labeling with 64Cu and found opposed results [31]. [ 64Cu]Oxo-DO3A-trastuzumab showed superior in vivo and in vitro characteristics compared to 64Cu-DOTA-trastuzumab. A comparison of in vivo results is very difficult due to the fact that pharmacokinetic properties are mainly driven by the antibody whereas the chelator has even less impact on biodistribution. Kiefer and co-workers also showed a higher stability in mouse serum and a higher tumor uptake for the Oxo-DO3A antibody than for the DOTA conjugated analogue [31]. These and our findings indicate that OxoDO3A may be a suitable chelator for labeling with 64Cu but not for labeling with 68Ga. 5. Conclusion Altogether, this study demonstrates that a variety of parameters are involved in the selection of the most successful chelating system for a 68Ga-labeled radiotracer. Despite, initial data like complex stability constants or even successfully used compounds (e.g., [ 68Ga] DOTA-TOC) suggest potential use of corresponding chelating systems, combination with the peptide of interest may change the in vitro and in vivo behavior making an a priori prediction difficult. Our test series,

Bettina Sarg, Division of Clinical Biochemistry, Innsbruck Medical University is acknowledged for carrying out the LC-MS analysis. We thank David A. Cheresh, The Scripps Research Institute, La Jolla, CA, for providing the human melanoma M21 and M21-L cells.

References [1] Weis SM, Cheresh DA. αv integrins in angiogenesis and cancer. Cold Spring Harb Perspect Med 2011;1:a006478. [2] Shojaei F. Anti-angiogenesis therapy in cancer: current challenges and future perspectives. Cancer Lett 2012;320:130-7. [3] Heidenreich R, Röcken M, Ghoreschi K. Angiogenesis drives psoriasis pathogenesis. Int J Exp Pathol 2009;90:232-48. [4] Sadeghi MM, Glover DK, Lanza GM, Fayad ZA, Johnson LL. Imaging atherosclerosis and vulnerable plaque. J Nucl Med 2010;51(Suppl.):S51-65. [5] Bishop GG, McPherson JA, Sanders JM, Hesselbacher SE, Feldman MJ, McNamara CA, et al. Selective αvβ3-receptor blockade reduces macrophage infiltration and restenosis after balloon angioplasty in the atherosclerotic rabbit. Circulation 2001;103:1906-11. [6] Storgard CM, Stupack DG, Jonczyk A, Goodman SL, Fox RI, Cheresh DA. Decreased angiogenesis and arthritic disease in rabbits treated with an αvβ3 antagonist [see comments]. J Clin Invest 1999;103:47-54. [7] Haubner R, Decristoforo C. Radiolabelled RGD peptides and peptidomimetics for tumour targeting. Front Biosci 2009;14:872-86. [8] Haubner R, Wester HJ, Weber WA, Mang C, Ziegler SI, Goodman SL, et al. Noninvasive imaging of αvβ3 integrin expression using 18F-labeled RGDcontaining glycopeptide and positron emission tomography. Cancer Res 2001; 61:1781-5. [9] Beer AJ, Grosu AL, Carlsen J, Kolk A, Sarbia M, Stangier I, et al. [18F]Galacto-RGD positron emission tomography for imaging of αvβ3 expression on the neovasculature in patients with squamous cell carcinoma of the head and neck. Clin Cancer Res 2007;13:6610-6. [10] Beer AJ, Lorenzen S, Metz S, Herrmann K, Watzlowik P, Wester HJ, et al. Comparison of integrin αvβ3 expression and glucose metabolism in primary and metastatic lesions in cancer patients: a PET study using 18F-galacto-RGD and 18 F-FDG. J Nucl Med 2008;49:22-9. [11] Beer AJ, Niemeyer M, Carlsen J, Sarbia M, Nahrig J, Watzlowik P, et al. Patterns of αvβ3 expression in primary and metastatic human breast cancer as shown by 18 F-Galacto-RGD PET. J Nucl Med 2008;49:255-9. [12] Gaertner FC, Kessler H, Wester HJ, Schwaiger M, Beer AJ. Radiolabelled RGD peptides for imaging and therapy. Eur J Nucl Med 2012;39(Suppl 1):S126-38. [13] Decristoforo C, Hernandez Gonzalez I, Carlsen J, Rupprich M, Huisman M, Virgolini I, et al. 68Ga- and 111In-labelled DOTA-RGD peptides for imaging of αvβ3 integrin expression. Eur J Nucl Med Mol Imaging 2008;35:1507-15. [14] Eisenwiener KP, Prata MI, Buschmann I, Zhang HW, Santos AC, Wenger S, et al. NODAGATOC, a new chelator-coupled somatostatin analogue labeled with [67/68Ga] and [111In] for SPECT, PET, and targeted therapeutic applications of somatostatin receptor (hsst2) expressing tumors. Bioconjug Chem 2002;13: 530-41. [15] Seifert S, Kunstler JU, Schiller E, Pietzsch HJ, Pawelke B, Bergmann R, et al. Novel procedures for preparing 99mTc(III) complexes with tetradentate/monodentate coordination of varying lipophilicity and adaptation to 188Re analogues. Bioconjug Chem 2004;15:856-63. [16] Knetsch PA, Petrik M, Griessinger CM, Rangger C, Fani M, Kesenheimer C, et al. [68Ga]NODAGA-RGD for imaging αvβ3 integrin expression. Eur J Nucl Med Mol Imaging 2011;38:1303-12. [17] Breeman WA, de Jong M, de Blois E, Bernard BF, Konijnenberg M, Krenning EP. Radiolabelling DOTA-peptides with 68Ga. Eur J Nucl Med Mol Imaging 2005;32: 478-85.

72

P.A. Knetsch et al. / Nuclear Medicine and Biology 40 (2013) 65–72

[18] Haubner R, Kuhnast B, Mang C, Weber WA, Kessler H, Wester HJ, et al. [18F] Galacto-RGD: synthesis, radiolabeling, metabolic stability, and radiation dose estimates. Bioconjug Chem 2004;15:61-9. [19] Decristoforo C, Santos I, Pietzsch HJ, Kuenstler JU, Duatti A, Smith CJ, et al. Comparison of in vitro and in vivo properties of [99mTc]cRGD peptides labeled using different novel Tc-cores. Q J Nucl Med Mol Imaging 2007;51:33-41. [20] Sun Y, Anderson CJ, Pajeau TS, Reichert DE, Hancock RD, Motekaitis RJ, et al. Indium (III) and gallium (III) complexes of bis(aminoethanethiol) ligands with different denticities: stabilities, molecular modeling, and in vivo behavior. J Med Chem 1996;39:458-70. [21] Dijkgraaf I, Yim CB, Franssen GM, Schuit RC, Luurtsema G, Liu S, et al. PET imaging of αvβ3 integrin expression in tumours with 68Ga-labelled mono-, di- and tetrameric RGD peptides. Eur J Nucl Med Mol Imaging 2011;38:128-37. [22] Govindaswamy N, Quarless DA, Koch SA. New amine trithiolate tripod ligand and its iron(II) and iron (III) complexes. J Am Chem Soc 1995;117:8468-9. [23] Motekaitis RJ, Martell AE, Koch SA, Hwang JW, Duncan AQ, Welch MJ. The gallium(III) and indium(III) complexes of Tris(2-mercaptobenzyl)amine and Tris(2-hydroxybenzyl)amine. Inorg Chem 1998;37:5902-11. [24] Cutler CS, Giron MC, Reichert DE, Snyder AZ, Herrero P, Anderson CJ, et al. Evaluation of gallium-68 tris(2-mercaptobenzyl)amine: a complex with brain and myocardial uptake. Nucl Med Biol 1999;26:305-16.

[25] Luyt LG, Katzenellenbogen JA. A trithiolate tripodal bifunctional ligand for the radiolabeling of peptides with gallium(III). Bioconjug Chem 2002;13:1140-5. [26] Kunstler JU, Bergmann R, Gniazdowska E, Koźmiński P, Walther M, Pietzsch HJ. Impact of functionalized coligands on the pharmacokinetics of 99mTcIII ‘4+1’ mixed-ligand complexes conjugated to bombesin. J Inorg Biochem 2011;105: 1383-90. [27] Noll B. Mixed-ligand 68Ga complexes with tetradentate/monodentate NS3/N (‘4 +1’) coordination: relation of structure with in vitro stability. In: Mazzi U, editor. Technetium, rhenium and other metals in chemistry and nuclear medicine. Padova: SGEditoriali; 2006. p. 227-8. [28] Delgado R, Quintino S, Teixeira M, Zhang A. Metal complexes of a 12-membered tetraaza macrocycle containing pyridine and N-carboxymethyl groups. J Chem Soc Dalton Trans 1997:55-64. [29] Fani M, Maecke HR. Radiopharmaceutical development of radiolabelled peptides. Eur J Nucl Med Mol Imaging 2012;39(Suppl 1):S11-30. [30] Ferreira CL, Lamsa E, Woods M, Duan Y, Fernando P, Bensimon C, et al. Evaluation of bifunctional chelates for the development of gallium-based radiopharmaceuticals. Bioconjug Chem 2010;21:531-6. [31] Ferreira CL, Yapp DT, Crisp S, Sutherland BW, Ng SW, Gleave M, et al. Comparison of bifunctional chelates for 64Cu antibody imaging. Eur J Nucl Med Mol Imaging 2010;37:2117-26.