F-Labeling, and - American Chemical Society

Bioconjugate Chem. 2008, 19, 1871–1879

1871

Synthesis, 18F-Labeling, and in Vitro and in ViWo Studies of Bombesin Peptides Modified with Silicon-Based Building Blocks Aileen Ho¨hne,† Linjing Mu,† Michael Honer,† P. August Schubiger,† Simon M. Ametamey,*,† Keith Graham,‡ Timo Stellfeld,‡ Sandra Borkowski,‡ Dietmar Berndorff,‡ Ulrich Klar,‡ Ulrike Voigtmann,‡ John E. Cyr,‡ Matthias Friebe,‡ Ludger Dinkelborg,‡ and Ananth Srinivasan‡ Center for Radiopharmaceutical Science of ETH, PSI and USZ, Department of Chemistry and Applied Biosciences, ETH Zurich, CH-8093 Zurich, Switzerland, and Bayer Schering Pharma AG, Global Drug Discovery, D-13342 Berlin, Germany. Received April 17, 2008; Revised Manuscript Received July 11, 2008

The gastrin-releasing peptide receptor (GRPr) is overexpressed on various human tumors. The goal of our study was the synthesis of new 18F-labeled bombesin analogues for the PET imaging of GRPr expression in prostate tumor using a silicon-based one-step n. c. a. radiolabeling method. The silicon-containing building blocks were efficiently coupled to the N-terminus of the peptides via solid-phase synthesis. Radiolabeling of the obtained peptide precursors proceeded smoothly under acidic conditions (34-85% conversion). Using the di-tert-butyl silyl building block as labeling moiety, products containing a hydrolytically stable 18F-label were obtained. In in Vitro receptor binding experiments 2-(4-(di-tert-butylfluorosilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-Ala-Val-NMeGlyHis-Sta-Leu-NH2 (4b, IC50 ) 22.9 nM) displayed a 12-fold higher binding affinity than 2-(4-(di-tertbutylfluorosilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-Ala-Val-Gly-His(3Me)-Sta-Leu-NH2 (3b, IC50 ) 276.6 nM), and 4b was therefore chosen for further evaluation. In Vitro and ex ViVo metabolite studies of [18F]4b showed no significant degradation. In biodistribution experiments, tumor uptake of [18F]4b was low and unspecific, whereas the GRPr-rich pancreas revealed a high and specific accumulation of the radiotracer. This study demonstrates the applicability of our silicon-based one-step n. c. a. radiolabeling method for the synthesis of new 18F-labeled bombesin derivatives. This innovative approach represents a general, straightforward access to radiolabeled peptides as PET imaging probes.

INTRODUCTION Radiolabeled peptides have been the focus of research over the past decades due to their potential as molecular imaging probes and therapeutic agents (1, 2). Among the peptides for tumor targeting, bombesin (BBN) and bombesin derivatives have received significant interest as they exhibit high affinity and selectivity for the gastrin-releasing peptide receptor (GRPr) (3, 4). The GRPr is overexpressed on various human tumors including prostate, breast, pancreatic, and small cell lung cancers (5). With respect to prostate cancer, a high incidence of GRP receptors was found in human normal prostate tissues, in contrast to their very low abundance in non-neoplastic prostatic tissues (6). Thus, radiolabeled bombesin peptides are potentially valuable tools for the discrimination of prostate cancers from prostate hyperplasia and prostatitis. Accordingly, a lot of research has gone into the development of optimized BBN derivatives for imaging and therapy involving a wide range of peptide sequences, labeling moieties, and radionuclides. Research efforts have been devoted in particular to the optimization of peptide sequences leading to improved binding affinity and metabolic stability. The C-terminus of bombesin is required for high-affinity binding and biological potency (7); therefore, usually the N-terminus of bombesin is modified in order to allow labeling with radioisotopes. A number * To whom correspondence should be addressed. (S.M.A.) Center for Radiopharmaceutical Science of ETH, PSI and USZ, ETH Ho¨nggerberg D-CHAB IPW HCI H427, Wolfgang-Pauli-Strasse 10, CH8093 Zurich, Switzerland. Tel.: +41 44 633 74 63. Fax: +41 44 633 13 67. E-mail address: [email protected]. † ETH Zurich. ‡ Bayer Schering Pharma AG.

of 99mTc- and 111In-labeled BBN analogues have been developed as diagnostic probes for single-photon emission computed tomography (SPECT) (8). For potential peptide receptor radiotherapy applications, 90Y-, 177Lu-, and 188Re-labeled BBN conjugates have been evaluated (9, 10). Despite the obvious advantage of being a more quantitative imaging method with better spatial resolution and image contrast compared to SPECT, PET imaging of GRPr expression has been less studied and labeling of bombesin conjugates has been limited to metallic PET nuclides such as 64Cu, 68Ga, and 86 Y (11, 12). Among the available PET nuclides, 18F has ideal characteristics for peptide receptor imaging studies regarding its half-life (109.7 min) and low β+ energy (0.64 MeV) (13). However, so far only one report on the 18F-labeling of bombesin analogues including their in Vitro and in ViVo evaluation has been published (14). 18F has been introduced into two bombesin derivates through the prosthetic labeling group N-succinimidyl4-[18F]fluorobenzoate (18F-SFB). This approach, however, requires a multistep reaction sequence and is therefore laborious. Very recently, Schirrmacher et al. described the two-step radiolabeling of a N-amino-oxy derivatized bombesin peptide using the labeling synthon p-(di-tert-butylfluorosilyl)benzaldehyde ([18F]SiFA-A) (15). Due to low precursor amounts, high specific activities could be obtained (225-680 GBq/µmol for [18F]SiFA-A), although the 18F-fluorination was achieved Via isotopic exchange. Data from the biological evaluation of this bombesin derivative are currently not available. We have developed new silicon-based building blocks for the one-step n. c. a. 18F-labeling of biomolecules by nucleophilic displacement of an alkoxy, hydroxy, or hydride leaving group (16, 17). These building blocks can be efficiently coupled to the N-terminus of peptides during solid-phase synthesis.

10.1021/bc800157h CCC: $40.75  2008 American Chemical Society Published on Web 08/28/2008

1872 Bioconjugate Chem., Vol. 19, No. 9, 2008

Ho¨hne et al.

Figure 1. Structures of bombesin derivatives.

Facile one-step radiolabeling of model compounds and tetrapeptides containing these building blocks could be demonstrated. An 18F-labeled di-tert-butyl silyl based tetrapeptide has been shown to be stable against defluorination under physiological conditions (17). Here, we report on the application of this one-step labeling protocol to the radiosynthesis of a series of 18F-labeled bombesin analogues (Figure 1). Moreover, we evaluated the analogue with the best in Vitro binding affinity for its metabolic stability and in ViVo distribution.

EXPERIMENTAL SECTION General. All chemicals unless otherwise stated were purchased from Sigma-Aldrich or Merck and used without further purification. The syntheses of 2,5-dioxopyrrolidin-1-yl 2-(4(hydroxydi-iso-propylsilyl)phenyl)acetate and 2-(4-(di-tert-butylsilyl)phenyl)acetic acid were previously reported (17). Peptide synthesis was carried out using Rink-Amide resin (0.68 mmol/ g) following standard 9-fluorenylmethyloxycarbonyl (Fmoc) strategy (18). All amino acid residues were, if not further specified, L-amino acid residues. FA01010 is (4R,5S)-4-amino-

5-methylheptanoic acid, Statine (Sta) is (3S,4S)-4-amino-3hydroxy-6-methylheptanoic acid, and Ava is 5-aminopentanoic acid. High-performance liquid chromatography (HPLC) analyses were performed using either a Hamilton PRP-1 (250 × 4.1 mm, 7 µm) or an ACE C18 column (50 × 4.6 mm, 3 µm) and the indicated conditions. Analytical HPLC systems used were Agilent 1100 systems with Gina or Agilent software and a Merck-Hitachi L6200A system, equipped with UV multiwavelength and Raytest Gabi Star detectors. Semipreparative HPLC purifications were carried out using a semipreparative ACE C18 column (250 × 10 mm, 5 µm) and the indicated conditions. Semipreparative HPLC systems used were an Agilent 1100 system equipped with UV multiwavelength and Raytest Gabi Star detector and a Merck-Hitachi L6200A system equipped with Knauer variable wavelength detector and Eberline radiation monitor. UV chromatograms were recorded at 218 or 230 nm. For the ex ViVo metabolite studies, an Ultraperformance liquid chromatography (UPLC) system from Waters with a Waters

18

F-Silyl Bombesin Peptides

Acquity UPLC BEH C18 column (2.1 × 50 mm, 1.7 µm) and an attached Berthold coincidence detector (FlowStar LB513) was used. All animal experiments were performed in compliance with the current version of the German law on the Protection of Animals. Peptide Synthesis. Fmoc-Deprotection (General Procedure). The resin-bound Fmoc peptide was treated with 20% piperidine in DMF (v/v) for 5 min. This step was repeated with a reaction time of 20 min. The resin was washed with DMF (2×), CH2Cl2 (2×), and DMF (2×). HBTU/HOBT Coupling (General Procedure). A solution of Fmoc-Xaa-OH (Xaa ) amino acid, 4 equiv), HBTU (O(benzotriazol-1-yl)-N,N,N,N-tetramethyluronium hexafluorophosphate, 4 equiv), HOBT (1-hydroxybenzotriazole, 4 equiv), DIPEA (N,N′-di-iso-propylethylamine, 4 equiv) in DMF was added to the resin-bound, free amine peptide and shaken for 90 min at room temperature. This step was repeated with a reaction time of 60 min. The resin was washed with DMF (2×), CH2Cl2 (2×), and DMF (2×). The peptides were typically prepared starting with 147 mg (100 µmol) of the resin. The amounts of reagents and building blocks in all subsequent reactions were calculated based on this amount. Synthesis of 3-(3-(Di-iso-propylhydroxysilyl)phenyl)acetylAVa-Gln-Trp-Ala-Val-Gly-His(3Me)-FA01010-Leu-NH2 (Precursor Compound 1a). The resin-bound, side chain protected peptide was prepared according to the general procedures described above. 2,5-Dioxopyrrolidin-1-yl 2-(4-(hydroxydi-isopropylsilyl)phenyl)acetate (54.5 mg, 150 µmol) was dissolved in DMF (1 mL). The resin-bound peptide (100 µmol) was suspended in this solution and the suspension was shaken for 12 h at ambient temperature. The resin was then filtered, washed with DMF (3 × 3 mL) and CH2Cl2 (3 × 3 mL), and dried in Vacuo. Subsequent treatment of the resin with 2 mL TFA/water/ tri-iso-propylsilane/phenol (85:5:5:5) afforded the crude, fully deprotected peptide, which was precipitated and washed with cold methyl tert-butyl ether. The crude peptide was dried in Vacuo, purified by preparative RP-HPLC, and lyophilized. The purified peptide (42 mg, 32%) was analyzed by HPLC-MS (m/z [M + H]+ calcd.: 1311.8, found: 1311.7). Procedure for the Syntheses of Di-tert-butylsilyl-Functionalized Peptides (Precursor Compounds 2a-4a). Representative procedure: The resin-bound, side chain protected peptide was prepared according to the general procedures described above. 2-(4-(Di-tert-butylsilyl)phenyl)acetic acid (55.7 mg, 200 µmol), HBTU (75.8 mg, 200 µmol), and HOBt (30.6 mg, 200 µmol) were dissolved in DMF (1.5 mL) and DIPEA (68 µL, 400 µmol) was added. The resin-bound peptide (100 µmol) was suspended in this solution and the suspension was shaken for 12 h at ambient temperature. The resin was then filtered, washed with DMF (3 × 3 mL) and CH2Cl2 (3 × 3 mL), and dried in Vacuo. Subsequent treatment of the resin with 2 mL TFA/water/triiso-propylsilane/phenol (85:5:5:5) afforded the crude, fully deprotected peptide, which was precipitated and washed with cold methyl tert-butyl ether. The crude peptide was dried in Vacuo, purified by preparative RP-HPLC, and lyophilized. The products were analyzed by HPLC-MS: 2-(4-(Di-tert-butylsilyl)phenyl)acetyl-Ava-Gln-Trp-Ala-ValNMeGly-His(3Me)-FA01010-Cpa-NH2 (2a, 10 mg, 7.5%): m/z [M + H]+ calcd.: 1335.8 found: 1335.9. 2-(4-(Di-tert-butylsilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-AlaVal-Gly-His(3Me)-Sta-Leu-NH2 (3a, 18.5 mg, 12%): m/z [(M + 2H)/2]+ calcd.: 748.5, found: 749.0. 2-(4-(Di-tert-butylsilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-AlaVal-NMeGly-His-Sta-Leu-NH2 (4a, 33 mg, 22%): m/z [(M + 2H)/2]+ calcd.: 748.5, found: 749.0.

Bioconjugate Chem., Vol. 19, No. 9, 2008 1873

Synthesis of Fluorinated Peptides (Reference Compounds 1b-4b). Representative procedure: The silane- or hydroxysilane-functionalized peptide (1a-4a, 3.3 µmol) was dissolved in THF (0.5 mL). This solution was added to a mixture of KF (0.8 mg, 13.1 µmol), K2.2.2. (Kryptofix 2.2.2., 4.9 mg, 13.1 µmol), and K2CO3 (0.9 mg, 6.6 µmol). Glacial acetic acid (2.2 µL, 39 µmol) was added and the resulting suspension was heated at 60 °C for 30 min. The crude mixture was directly subjected to preparative RP-HPLC, and the purified product was lyophilized. The products were analyzed by HPLC-MS: 3-(3-(Di-iso-propylfluorosilyl)phenyl)acetyl-Ava-Gln-TrpAla-Val-Gly-His(3Me)-FA01010-Leu-NH2 (1b, 2 mg, 40%): m/z [M + H]+ calcd.: 1313.8, found: 1313.5. 2-(4-(Di-tert-butylfluorosilyl)phenyl)acetyl-Ava-Gln-Trp-AlaVal-NMeGly-His(3Me)-FA01010-Cpa-NH2 (2b, 2.5 mg, 50%): m/z [M + H]+ calcd.: 1353.8 found: 1353.8. 2-(4-(Di-tert-butylfluorosilyl)phenyl)acetyl-Arg-Ava-Gln-TrpAla-Val-Gly-His(3Me)-Sta-Leu-NH2 (3b, 3 mg, 59%): m/z [(M + 2H)/2]+ calcd.: 757.5, found: 758.0. 2-(4-(Di-tert-butylfluorosilyl)phenyl)acetyl-Arg-Ava-Gln-TrpAla-Val-NMeGly-His-Sta-Leu-NH2 (4b, 2 mg, 40%): m/z [(M + 2H)/2]+ calcd.: 757.5, found: 758.0. Radiolabeling and Stability Studies. No-carrier-added [18F]fluoride was produced via the 18O(p,n)18F nuclear reaction by irradiation of enriched [18O]H2O. [18F]Fluoride was trapped on an anion-exchange resin cartridge (Sep-Pak QMA light, Waters). The cartridge was eluted with a solution of Kryptofix (K2.2.2., 5 mg) and potassium carbonate (1 mg) in H2O (0.5 mL) and CH3CN (1.0-1.5 mL). Solvents were removed by heating at 100-140 °C for 20-30 min applying a gentle stream of nitrogen. During this time, CH3CN (3 × 1 mL) was added and evaporated to give the dry K[18F]F/K2.2.2. complex. The peak of the 18F-labeled products was confirmed by comparison with the HPLC retention time of their nonradioactive reference molecule or by coinjection. 3-(3-(Di-iso-propyl[18F]fluorosilyl)phenyl)acetyl-AVa-GlnTrp-Ala-Val-Gly-His(3Me)-FA01010-Leu-NH2 ([18F]1b). A solution of 1a (2.0 mg) and glacial acetic acid (3 µL) in anhydrous DMSO (150 µL) was added to the dry K[18F]F/K2.2.2. complex. After heating at 90 °C for 15-30 min, an aliquot of the reaction mixture was diluted with CH3CN (1 mL). The crude reaction mixture was analyzed using an analytical HPLC (PRP-1; isocratic 10 mM K2HPO4 in CH3CN/H2O (7:3)/10 mM K2HPO4 in H2O 55:45, 1.0 mL/min). Hydrolytic stability of the 18Flabel was tested by addition of water (9 mL) to the CH3CNdiluted aliquot of the reaction mixture and analysis of this mixture by HPLC at different time points. 2-(4-(Di-tert-butyl[18F]fluorosilyl)phenyl)acetyl-AVa-Gln-TrpAla-Val-NMeGly-His(3Me)-FA01010-Cpa-NH2 ([18F]2b). A solution of 2a (1.0-1.5 mg) in anhydrous DMSO (150 µL) was added to the dry K[18F]F/K2.2.2. complex. After heating at 70 °C for 30 min or 90 °C for 15 min, an aliquot of the reaction mixture was diluted with CH3CN (1 mL). The crude reaction mixture was analyzed using an analytical HPLC (PRP-1; isocratic 10 mM K2HPO4 in CH3CN/H2O (7:3)/10 mM K2HPO4 in H2O 70:30, 0.8 mL/min). The reaction mixture was diluted with H2O (8-9 mL) and slowly eluted through a Chromafix C18 (s) cartridge. The product was eluted with ethanol (1.5-2 mL) and analyzed by HPLC. Hydrolytic stability of [18F]2b was tested by addition of PBS to the ethanolic solution of the product and analyses of this mixture by HPLC at different time points. 2-(4-(Di-tert-butyl[18F]fluorosilyl)phenyl)acetyl-Arg-AVa-GlnTrp-Ala-Val-Gly-His(3Me)-Sta-Leu-NH2 ([18F]3b). A solution of 3a (2.0 mg) and glacial acetic acid (5 µL) in anhydrous DMSO (150 µL) was added to the dry K[18F]F/K2.2.2. complex. After heating at 110 °C for 20 min, an aliquot of the reaction


mixture was diluted with CH3CN (1 mL). The crude reaction mixture was analyzed using an analytical HPLC (ACE C18; gradient 10 mM K2HPO4 in CH3CN/H2O (7:3)/10 mM K2HPO4 in H2O 5:95-95:5 in 7 min, 2.0 mL/min). The reaction mixture was diluted with HPLC eluent (CH3CN/H2O + 0.1% TFA 45: 55, 4.0 mL). This solution was injected into a semipreparative HPLC (ACE C18; isocratic CH3CN/H2O + 0.1% TFA 45:55, 3.0 mL/min) and the product peak was collected. 2-(4-(Di-tert-butyl[18F]fluorosilyl)phenyl)acetyl-Arg-AVa-GlnTrp-Ala-Val-NMeGly-His-Sta-Leu-NH2 ([18F]4b). A solution of 4a (2.0 mg) and glacial acetic acid (5 µL) in anhydrous DMSO (150 µL) was added to the dry K[18F]F/K2.2.2. complex and heated at 110 °C for 20 min. An aliquot of the crude reaction mixture was analyzed using an analytical HPLC (ACE C18; gradient CH3CN/H2O + 0.1% TFA 5:95-95:5 in 10 min, 1.0 mL/min) either before or after addition of 4.0-4.5 mL HPLC eluent (CH3CN/H2O + 0.1% TFA 42:58) into the reaction vial. The diluted reaction mixture was injected into a semipreparative HPLC (ACE C18; isocratic CH3CN/H2O + 0.1% TFA 42:58, 3.0 mL/min) and the product peak was collected. The product fraction was diluted with water (20-45 mL) and immobilized on a C18 cartridge (Sep-Pak light C18, Waters, or Chromafix C18 (s), Machery-Nagel). After washing with water (2 × 10 mL), [18F]4b was eluted with ethanol or 5% 0.1 N HCl in ethanol (2 mL). The solvent was evaporated at 90 °C. For in Vitro applications, [18F]4b was reconstituted in 0.02 M phosphate buffer or phosphate buffered saline (PBS) containing max. 5% ethanol. For ex ViVo biodistribution experiments, the radiotracer was reconstituted in 0.9% NaCl containing max 5% ethanol; for ex ViVo metabolite studies, dried [18F]4b was dissolved in PEG-200/H2O (1:1). Log D Measurement. The experimental determination of distribution coefficient of [18F]4b was performed in 1-octanol and 0.06 M phosphate buffer at a pH of 7.4. The two phases were presaturated with each other. 1-Octanol (0.5 mL) and phosphate buffer (0.5 mL) were pipetted into five tubes (1.5 mL) containing of [18F]4b (5 µL). The tubes were shaken for 15 min at room temperature and centrifuged (3 min, 5000 U/min). From each tube, 50 µL of each phase was transferred into an empty tube for counting in a γ-counter. In Vitro Receptor Binding Assay. The receptor affinity of two BBN analogues, 3b and 4b, was determined in quadruplicate in a scintillation proximity assay (SPA) using human bombesin 2 receptor membranes (PerkinElmer). The membranes were suspended in 20 mM HEPES, 3 mM MgCl2, 1 mM EDTA, and 10% sucrose (pH 7.4) and kept on ice until use. Assay buffer was 50 mM Tris (pH 7.2, HCl, Roth), 5 mM MgCl2, 1 mM EGTA, 0.3% polyethylenemine solution, protease inhibitor (Roche Diagnostics GmbH, 1 tablet/50 mL). For competition binding studies, membranes solution (106 µg/ml, 10 µL, 1.1 µg protein per well), suspended beads (agglutinin coupled SPA beads, Amersham Bioscience, 40 mg/mL, 10 µL), [125I]-Tyr4bombesin (PerkinElmer, specific activity: 81.4 TBq/mmol, 0.50 nM, 20 µL), and increasing concentrations of test compounds (8 pM-1500 nM, 10 µL) were added to well plates to give a final volume of 50 µL. For determination of the total binding, the compound was replaced by the corresponding volume of assay buffer. Unspecific binding was measured using 1000× excess of bombesin. After incubation for 2 h at room temperature, 50 µL of assay buffer was added to the wells. The plates were centrifuged (1700 rpm) and counted in a Top-counter (PerkinElmer). IC50 values were calculated using GraFit 5 software (Erithacus software, version 5.0.6). Ki values were calculated using the Cheng-Prusoff equation {Ki ) IC50/(1 + [L]/Kd)}. In Vitro Plasma Stability. The degradation of [18F]4b was investigated in mouse and human plasma from healthy donors.

Ho¨hne et al.

PBS was used as a control. [18F]4b was incubated for different time periods (from 0 to 2 h) at 37 °C in PBS, and human and mouse plasma. After incubation, plasma proteins were precipitated with CH3CN + 0.1% TFA and samples were centrifuged (10 min, 13 000 rpm) at 4 °C. The PBS controls were diluted with the same volume of CH3CN + 0.1% TFA. The supernatants and PBS controls were analyzed by HPLC (ACE C18; gradient CH3CN/H2O + 0.1% TFA 5:95-95:5 in 10 min, 1.0 mL/min). The percentage of intact peptide was determined at different incubation times. Ex ViWo Metabolite Studies. Mice were injected intravenously with 120-200 MBq of [18F]4b). Animals were sacrificed at 10 and 30 min p.i. (one animal per time point), and blood and urine were collected. An aliquot of plasma and urine was mixed with CH3CN for protein precipitation. The samples were centrifuged (5000 rpm, Megafuge R, Heraeus) and the supernatants were analyzed by UPLC (Waters Acquity UPLC BEH C18; gradient CH3CN/H2O + 0.1% TFA 5:95-95:5 in 3 min, 0.8 mL/min). Biodistribution. The biodistribution experiments were performed using nude mice bearing a human prostate tumor (PC3). For the induction of tumor xenografts, PC-3 cells (2 × 106 cells/mouse) were injected subcutaneously and allowed to grow for four weeks. Animals were injected intravenously with 150 kBq of [18F]4b (100 µL, 2.4 fmol). The animals were sacrificed at different postinjection times (from 0.5 to 4 h, n ) 3 for each time point). Organs and tissues of interest were collected and weighed. The amount of radioactivity was determined in the γ-counter to calculate the uptake (% injected dose per g tissue). In addition, a group of 3 mice received 100 µg of GRP coinjected with the radiolabeled compound and was sacrificed at 1 h postinjection to determine unspecific uptake.

RESULTS Peptide Synthesis. Peptide synthesis was carried out using Rink Amide resin following standard Fmoc strategy (18). Coupling of the resin-bound peptide with the activated ester of 2-(4-(hydroxydi-iso-propylsilyl)phenyl)acetic acid was achieved directly in DMF at room temperature, whereas the conjugation with 2-(4-(di-tert-butylsilyl)phenyl)acetic acid required the use of a coupling reagent system (HBTU/HOBt/DIPEA). The 19Freference compounds were obtained by direct fluorination of the appropriate precusors with KF in the presence of Kryptofix K2.2.2. and glacial acetic acid. Representative HPLC-MS chromatograms of a peptide precusor and reference are shown in Figure 2. Radiolabeling and Hydrolytic Stability Studies. The direct 18 F-fluorination protocol developed previously in our laboratory for the radiolabeling of model compounds and tetrapeptides was tested on a series of BBN peptide precursors. To demonstrate the general applicability of the method on targeting peptides, the radiolabeling of a di-iso-propylsilyl building block containing a BBN derivative was investigated. A reaction time of 15 min at 90 °C led to an 18F-incorporation of 4% (Table 1, entry 1). However, after heating for another 15 min at the same temperature,18F-incorporation increased to 34% (entry 2). We verified the hydrolytic stability of the 18F-label after the addition of water to an aliquot of the reaction mixture followed by HPLC analysis at different time points. Defluorination of the di-isopropylsilyl building block was observed: the percentage of [18F]1b in radiochromatograms decreased to 37% and 13% of the original value within 0.5 and 1 h, respectively. Subsequently, the radiolabeling of a BBN analogue containing the most stable di-tert-butylsilyl building block was studied. The applied nonacidic conditions gave a rather poor 18Fincorporation (entries 3 and 4, Table 1). As the radiolabel was expected to be stable under aqueous conditions, the product was

18



Figure 2. Representative HPLC-MS chromatograms of silyl-functionalized peptides: (a) 2-(4-(di-tert-butylsilyl)phenyl)acetyl-Arg-Ava-Gln-TrpAla-Val-NMeGly-His-Sta-Leu-NH2 (4a, peak 3 at 1.1 min; peak 1 and 2: system peaks) and (b) 2-(4-(di-tert-butylfluorosilyl)phenyl)acetyl-ArgAva-Gln-Trp-Ala-Val-NMeGly-His-Sta-Leu-NH2 (4b, peak 3 at 1.08 min; peak 1 and 2: system peaks). Table 1. Radiolabeling of Bombesin Derivatives precursor acetic amount acid entry precursor (mg) (µL) T (°C) t (min) 1 2 3 4 5 6 7

1a 1a 2a 2a 3a 4a 4a

2.0 2.0 1.0 1.5 2.0 2.0 2.0

3 3 0 0 5 5 5

90 90 70 90 110 110 110

15 30 30 15 20 20 20

18

F-incorporation (% HPLC)

4 34 15 19 85 73.5 ( 1.8% a (n ) 2) 31.0 ( 2.0% b (n ) 2)

a Aliquot directly removed from the reaction mixture. b Aliquot taken after dilution of the reaction mixture with HPLC eluent.

purified by solid-phase extraction (SPE) to avoid any influence of the other components in the reaction mixture during the hydrolytic stability study. [18F]2b was obtained in 69% radiochemical purity. A minor amount of 18F (6%) and side products (25%) could not be removed by this simple purification method. The stability test of [18F]2b in PBS showed that no defluorination occurred within 2 h. Radiolabeling of 3a was achieved under acidic conditions (entry 5, Table 1). Starting from 257 MBq 18F, 55 MBq of [18F]3b (radiochemical purity >99%) was obtained by semipreparative HPLC (30% d. c. RCY). 4a was labeled under the same conditions. If an aliquot was removed from the reaction mixture directly and analyzed by HPLC, the radiochromato-

grams showed 73.5 ( 1.8% product (entry 6). If an aliquot of the reaction mixture was taken after dilution with HPLC eluent, the percentage of [18F]4b measured by HPLC was 31.0 ( 2.0% (entry 7). This is probably an indication that not all of the K[18F]F/K2.2.2. complex was dissolved from the vessel walls into the reaction mixture, and therefore only a certain amount of the K[18F]F/K2.2.2. complex was available for the nucleophilic substitution reaction. HPLC purification and formulation gave [18F]4b in 13.1 ( 3.3% radiochemical yield (decaycorrected, n ) 4) and >95% radiochemical purity (Figure 4a) within 115-155 min. In a typical experiment, 2.78 GBq 18F provided 191 MBq of [18F]4b with a specific activity of 62 GBq/µmol (EOS). The distribution coefficient was determined in an equal volume mixture of 1-octanol and 0.06 M phosphate buffer (pH ) 7.4). The log D value was 1.3 ( 0.1 (n ) 5). In Vitro Receptor Binding Assay. The IC50 (Ki) values determined for 3 and 4 were 267.7 (69.7) and 22.9 (5.76) nM, respectively (Figure 3). Due to the higher binding affinity, [18F]4b was further evaluated in Vitro and in ViVo. In Vitro and ex ViVo Stability Studies. Metabolite studies did not show any significant degradation of [18F]4b in PBS or in mouse or human plasma within 2 h (Figure 4b-d). Also, after injection into mice, no degradation products of [18F]4b were observed in blood at 10 and 30 min postinjection. (Figure 5a-c). Metabolite analyses of urine revealed intact labeled peptide and [18F]-fluoride as a metabolite.


Ho¨hne et al.

Figure 3. Competitive binding assays of (a) 2-(4-(di-tert-butylfluorosilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-Ala-Val-Gly-His(3Me)-Sta-Leu-NH2 (3b) and (b) 2-(4-(di-tert-butylfluorosilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-Ala-Val-NMeGly-His-Sta-Leu-NH2 (4b) versus [125I]-Tyr 4-bombesin on human bombesin 2 receptor membranes in a scintillation proximity assay.

Figure 4. Radio-HPLC profiles of 2-(4-(di-tert-butyl[18F]fluorosilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-Ala-Val-NMeGly-His-Sta-Leu-NH2 ([18F]4b): (a) purified compound; (b) after 2 h in PBS; (c) after 2 h in human plasma; and (d) after 2 h in mouse plasma.

Biodistribution. The biodistribution of [18F]4b was evaluated in nude mice bearing subcutaneous PC-3 tumors. The results are shown in Table 2. Thirty minutes after injection, tumor uptake was 0.63 ( 0.13%ID/g. It essentially remained steady over the measured period of time and was not reduced by a blocking dose of bombesin. In contrast, uptake in the GRPrrich pancreas was high (between 5.41 ( 0.12%ID/g 0.5 h p.i. and 3.93 ( 1.11%ID/g 4 h p.i.) and specific (73% uptake reduction in the blocking group). Bone uptake increased from 1.35 ( 0.47%ID/g 0.5 h p.i. to 5.14 ( 2.71%ID/g 4 h p.i. High

accumulation of the radiotracer was found in liver, gallbladder, and intestine.

DISCUSSION Prostate tumors are characterized by low metabolic activity, which impairs the use of the most common, unspecific accumulating PET agents (e.g., FDG, 18F/11C-choline, 11C-acetate) for the imaging of prostate cancer (19). A PET tracer with high sensitivity and specificity is needed for the reliable diagnosis of lymph node involvement of prostate cancer at an early stage (first diagnosis) and for the clear discrimination between local

18



Figure 5. Radio-UPLC profiles of 2-(4-(di-tert-butyl[18F]fluorosilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-Ala-Val-NMeGly-His-Sta-Leu-NH2 ([18F]4b): (a) purified compound; (b) in blood 30 min postinjection into mouse; (c) in urine 30 min postinjection into mouse. Table 2. Biodistribution of 2-(4-(Di-tert-butyl[18F]fluorosilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-Ala-Val-NMeGly-His-Sta-Leu-NH2 ([18F]4b) in Nude Mice with PC-3 Xenografts at Different Times p.i.a

a

tissue

0.5 h

1.0 h

1.0 hb

2.0 h

4.0 h

blood kidney liver spleen gallbladder intestine bone pancreas tumor

0.95 ( 0.06 2.76 ( 0.27 13.56 ( 0.43 1.48 ( 1.50 177.65 ( 21.60 12.79 ( 2.96 1.35 ( 0.47 5.41 ( 0.12 0.63 ( 0.13

0.48 ( 0.01 1.96 ( 0.08 7.03 ( 1.69 0.71 ( 0.24 259.02 ( 129.67 18.94 ( 6.12 2.49 ( 0.08 4.67 ( 0.71 0.63 ( 0.03

1.04 ( 0.14 2.18 ( 0.36 22.30 ( 2.85 4.38 ( 1.44 118.89 ( 161.64 9.05 ( 0.44 2.58 ( 0.32 1.25 ( 0.55 0.61 ( 0.45

0.42 ( 0.04 1.73 ( 0.24 4.40 ( 1.04 0.46 ( 0.06 146.40 ( 126.00 15.74 ( 3.78 2.67 ( 0.09 4.08 ( 0.67 0.40 ( 0.05

0.37 ( 0.11 1.36 ( 0.19 3.55 ( 0.83 0.81 ( 0.13 146.43 ( 117.44 14.57 ( 2.89 5.14 ( 2.71 3.93 ( 1.11 0.58 ( 0.18

Results are expressed as %ID/g ( SD (n ) 3). b Blockade study: animals received 100 µg of GRP coinjected with the radiolabeled compound.

recurrence and distant metastases in the case of prostate-specific antigen (PSA) level increase after radical prostectomy, since both cases have different therapeutic implications. Peptides provide excellent characteristics for PET imaging, as they have fast and specific targeting features. They are rapidly cleared from the body, usually mainly via the urinary pathway (1). In particular, peptides targeting G-protein coupled receptors are effective in tumor accumulation because the cells internalize them. The gastrin-releasing peptide receptor (GRPr), which is overexpressed on prostatic tumors, belongs to this class of receptors. We have synthesized new BBN analogues (Figure 1) based on the amino acid sequence 7-14 (Gln7-Trp8-Ala9-Val10-Gly11His12-Leu13-Met14-NH2). In order to increase the metabolic stability of this fragment and to allow the attachment of the labeling moiety while retaining the binding affinity, a number of changes were introduced into the molecule. Met14 was replaced by Leu for stabilization against aminopeptidase 3.4.11.1, and Leu13 was replaced by a non-natural amino acid (FA01010 or Sta) to prevent cleavage by neutral endopeptidase 3.4.24.11 (20). Further modifications were the introduction of the methylated versions of His12 and/or Gly11 and the insertion of a

spacer, -Ava- or -Ava-Arg-, between the silicon-based building block and the binding sequence in order to avoid interference of the labeling moiety with the receptor binding C-terminus of the peptide. The new BBN analogues contained either a di-isopropyl or di-tert-butyl silyl building block as labeling moiety. The applicability of the direct labeling method for targeting peptides was successfully demonstrated on a di-iso-propyl silyl building block containing BBN derivative 1a. As expected from previous studies (17), defluorination of [18F]1b was observed under aqueous conditions. The radiolabeling of BBN analogues modified with the di-tert-butyl silyl building block proceeded smoothly under acidic conditions (72-85% conversion), whereas nonacidic conditions gave relatively low yields of 18Fincorporation (15-19%). A preliminary study of the hydrolytic stability of one of the di-tert-butyl silyl substituted 18F-labeled peptides showed no defluorination under aqueous conditions. In Vitro receptor binding studies verified a rather poor binding affinity for peptide 3b (Figure 3, IC50 ) 276.6 nM). Peptide 4b, however, displayed a 12-fold higher affinity (IC50 ) 22.9 nM) which is in the same range as the affinities of previously reported 18F-labeled bombesin analogues (IC50 ) 5.3 nM for FB-[Lys3]BBN and IC50 ) 48.7 nM for FB-Aca-BBN(7-14))


(14). [18F]4b was therefore chosen for in Vitro and ex ViVo stability tests. No significant degradation products of [18F]4b were detected in PBS, or mouse or human plasma within 2 h (Figure 4b-d). Compound [18F]4b also showed high in ViVo stability in the blood of mice (Figure 5b) demonstrating successful stabilization of the peptide sequence against degradation by endogenous enzymes. Biodistribution experiments of [18F]4b showed low and unspecific tumor uptake. In contrast, the GRPr-rich mouse pancreas revealed a high and specific accumulation of the radiotracer. This result may be explained by the higher receptor densities in pancreas compared to tumor and/or by the better accessibility of the GRPr-expressing pancreatic acini cells in comparison to the weakly vascularized and potentially necrotic PC-3 tumors. Bone uptake gradually increased over time, suggesting a slow defluorination (i.e., silanol formation) of the radiotracer. Due to the lipophilic character of [18F]4b, it exhibited a mainly hepatobiliary clearance as suggested by very strong signals in the liver, gallbladder, and intestine. Significant accumulation of radioactivity in the abdominal region can impair the effective imaging of tumors or metastatic lesions in this area of the body. Assuming that the lipophilicity of the entire molecule is decisive for its in ViVo behavior (21), a reduction of the overall lipophilicity of the radiolabeled peptide may lead to new bombesin derivatives with improved pharmacokinetic profile. An indication of the superiority of less lipophilic peptides is also the prolonged retention of radiometallabeled BBNs in PC-3 cells, which is most likely due to the lack of cell permeability of the hydrophilic macrocyclic conjugate (9, 22, 23). Furthermore, intracellular cleavage of the hydrophilic macrocycle from the peptide might lead to trapping of the radiometal inside the cells. On the opposite, 18F-labeled peptides contain inherently a lipophilic labeling moiety. Consequently, 18F-labeled BBNs, after GRPr-mediated internalization, are more amenable to penetration in and out of the cells due to the relatively lipophilic character of both the intact peptide and its radioactive metabolites. Zhang et al. found in biodistribution studies that the tumor uptake of 18F-FB-[Lys3]BBN (IC50 ) 5.3 nM) is relatively high (5.94 ( 0.78%ID/g at 60 min p.i.) compared to another derivative with lower binding affinity (18F-FB-Aca-BBN(7-14), IC50 ) 48.7 nM, 0.43 ( 0.18%ID/g at 60 min p.i.) (14). Hence, a high affinity may partly compensate for the comparatively high lipophilicity of 18F-labeled BBNs. Besides their hydrophilicity, the subnanomolar IC50 values of some of the radiometal-labeled BBN derivatives known in the literature (3, 4) are probably another factor for their favorable in ViVo characteristics. Thus, our future efforts will be directed toward the synthesis of both more potent and more hydrophilic 18F-silyl bombesin peptides for GRPr-targeted PET imaging.

CONCLUSION This study demonstrates the applicability of our silicon-based one-step n. c. a. radiolabeling method for the synthesis of 18Flabeled bombesin derivatives. Using the di-tert-butyl silyl building block as labeling moiety, products containing a hydrolytically stable 18F label were obtained. Biodistribution experiments of a di-tert-butyl silyl substituted 18F-labeled bombesin peptide revealed, besides substantial hepatobiliary clearance, low and unspecific uptake in GRPr-positive tumors. To enhance their suitability for the imaging of prostate tumors, optimization efforts toward a reduced overall lipophilicity and higher potency of our 18F-silyl BBNs will be made. In conclusion, we believe that our direct labeling method is transferable to targeting peptides in general and will allow straightforward access to novel PET imaging probes.

Ho¨hne et al.

ACKNOWLEDGMENT We gratefully acknowledge Cindy Fischer and Claudia Keller for some experimental support, Dr. Thomas Nauser and Mathias Nobst for technical support, and Marion Kuzora, Bodo Ro¨hr, and Frank Kuczynski for technical assistance.

LITERATURE CITED (1) Okarvi, S. M. (2004) Peptide-based radiopharmaceuticals: future tools for diagnostic imaging of cancers and other diseases. Med. Chem. ReV. 24, 357–397. (2) Reubi, J. C. (2003) Peptide receptors as molecular targets for cancer diagnosis and therapy. Endocr. ReV. 24, 389–427. (3) Smith, C. J., Volkert, W. A., and Hoffman, T. J. (2005) Radiolabeled peptide conjugates for targeting of the bombesin receptor superfamily subtypes. Nucl. Med. Biol. 32, 733–740. (4) Smith, C. J., Volkert, W. A., and Hoffman, T. J. (2003) Gastrin releasing peptide (GRP) receptor targeted radiopharmaceuticals: a concise update. Nucl. Med. Biol. 30, 861–868. (5) Patel, O., Shulkes, A., and Baldwin, G. S. (2006) Gastrinreleasing peptide and cancer. Biochim. Biophys. Acta 1766, 23– 41. (6) Markwalder, R., and Reubi, J. C. (1999) Gastrin-releasing peptide receptors in the human prostate: relation to neoplastic transformation. Cancer Res. 59, 1152–1159. (7) Moody, T. W., Crawley, J. N., and Jensen, R. T. (1982) Pharmacology and neurochemistry of bombesin-like peptides. Peptides 3, 559–563. (8) Varvarigou, A., Bouziotis, P., Zikos, C., Scopinaro, F., and De Vincentis, G. (2004) Gastrin-releasing peptide (GRP) analogues for cancer imaging. Cancer Biother. Radiopharm. 19, 219–229. (9) Zhang, H., Chen, J., Waldherr, C., Hinni, K., Waser, B., Reubi, J. C., and Ma¨cke, H. R. (2004) Synthesis and evaluation of bombesin derivatives on the basis of pan-bombesin peptides labeled with indium-111, lutetium-177, and yttrium-90 for targeting bombesin receptor-expressing tumors. Cancer Res. 64, 6707–6715. (10) Smith, C. J., Sieckman, G. L., Owen, N. K., Hayes, D. L., Mazuru, D. G., Volkert, W. A., and Hoffman, T. J. (2003) Radiochemical investigation of [188Re(H2O)(CO)3-diaminoproprionic acid-SSS-bombesin(7-14)NH2]: synthesis, radiolabeling and in vitro/in vivo GRP receptor targeting studies. Anticancer Res. 23, 63–70. (11) Biddlecombe, G. B., Rogers, B. E., de Visser, M., Parry, J. J., de Jong, M., Erion, J. L., and Lewis, J. S. (2007) Molecular imaging of gastrin-releasing peptide receptor-positive tumors in mice using 64Cu- and 86Y-DOTA-(Pro 1,Tyr 4)-bombesin(114). Bioconjugate Chem. 18, 724–730. (12) Schuhmacher, J., Zhang, H., Doll, J., Ma¨cke, H. R., Matys, R., Hauser, H., Henze, M., Haberkorn, U., and Eisenhut, M. (2005) GRP receptor-targeted PET of a rat pancreas carcinoma xenograft in nude mice with a 68Ga-labeled bombesin(6-14) analog. J. Nucl. Med. 46, 691–699. (13) Okarvi, S. M. (2001) Recent progress in fluorine-18 labelled peptide radiopharmaceuticals. Eur. J. Nucl. Med. 28, 929–938. (14) Zhang, X., Cai, W., Cao, F., Schreibmann, E., Wu, Y., Wu, J. C., Xing, L., and Chen, X. (2006) 18F-labeled bombesin analogs for tageting GRP receptor-expressing prostate cancer. J. Nucl. Med. 47, 492–501. (15) Schirrmacher, E., Wa¨ngler, B., Cypryk, M., Bradtmo¨ller, G., Scha¨fer, M., Eisenhut, M., Jurkschat, K., and Schirrmacher, R. (2007) Synthesis of p-(di-tert-butyl[18F]fluorosilyl)benzaldehyde ([18F]SiFA-A) with high specific activity by isotopic exchange: a convenient labeling synthon for the 18F-labeling of N-aminooxy derivatized peptides. Bioconjugate Chem. 18, 2085–2089. (16) Ho¨hne, A., Mu, L., Schubiger, P. A., Ametamey, S. M., Graham, K., Cyr, J. E., Dinkelborg, L., Stellfeld, T., Srinivasan, A., Voigtmann, U., and Klar, U. (2007) Development of new silicon-based building blocks 18F-labeling of biomolecules. J. Labelled Compd. Radiopharm. 50, S6.

18


(17) Mu, L., Ho¨hne, A., Schubiger, P. A., Ametamey, S. M., Graham, K., Cyr, J. E., Dinkelborg, L., Stellfeld, T., Srinivasan, A., Voigtmann, U., Klar, U. (2008) Silicon-based building blocks for one-step 18F-radiolabeling of peptides for PET imaging. Angew Chem., Int. Ed. 47, 4922–4925. (18) Fields, G. B., and Noble, R. L. (1990) Solid phase peptide synthesis utilizing 9- fluorenylmethoxycarbonyl amino acids. Int. J. Pept. Protein Res. 35, 161–214. (19) Fanti, S., Nanni, C., Ambrosini, V., Gross, M. D., Rubello, D., and Farsad, M. (2007) PET in genitourinary tract cancers. Q. J. Nucl. Med. Mol. Imaging 51, 260–271. (20) Davis, T. P., Crowell, S., Taylor, J., Clark, D. L., Coy, D., Staley, J., and Moody, T. W. (1992) Metabolic stability and tumor inhibition of bombesin/GRP receptor antagonists. Peptides 13, 401–407. (21) Smith, C. J., Gali, H., Sieckman, G. L., Higginbotham, C., Volkert, W. A., and Hoffman, T. J. (2003) Radiochemical

Bioconjugate Chem., Vol. 19, No. 9, 2008 1879 investigations of 99mTc-N3S-X-BBN 7-14 NH2: An in vitro/in vivo structure-activity relationship study where X ) 0-, 3-, 5-, 8-, and 11-carbon tethering moieties. Bioconjugate Chem. 14, 93–102. (22) Rogers, B. E., Bigott, H. M., McCarthy, D. W., Della Manna, D., Kim, J., Sharp, T. L., and Welch, M. J. (2003) MicroPET imaging of a gastrin-releasing peptide receptor-positive tumor in a mouse model of human prostate cancer using a 64Cu-labeled bombesin analogue. Bioconjugate Chem. 14, 756–763. (23) Scopinaro, F., De Vincentis, G., Corazziari, E., Massa, R., Osti, M., Pallotta, N., Covotta, A., Remediani, S., Di Paolo, M., Monteleone, F., and Varvarigou, A. (2004) Detection of colon cancer with 99m Tc-labeled bombesin derivative (99mTc-leu13BN1). Cancer Biother. Radiopharm. 19, 245–252. BC800157H

F-Labeling, and - American Chemical Society

Recommend Documents