Available online at www.sciencedirect.com

gy 36 (2009) 613–622www.elsevier.com/locate/nucmedbio

Nuclear Medicine and Biolo

Synthesis, radiolabeling, in vitro and in vivo evaluation of [18F]-FPECMOas a positron emission tomography radioligand for imaging the

metabotropic glutamate receptor subtype 5Christophe Lucatelli, Michael Honer, Jean-Frédéric Salazar, Tobias L. Ross,

P. August Schubiger, Simon M. Ametamey⁎

Center for Radiopharmaceutical Science of ETH, PSI and USZ, 8093 Zurich, SwitzerlandDepartment of Chemistry and Applied Biosciences, ETH Zurich, 8093 Zurich, Switzerland

Received 12 December 2008; received in revised form 5 March 2009; accepted 16 March 2009

Abstract

Introduction: [18F]-(E)-3-((6-Fluoropyridin-2-yl)ethynyl)cyclohex-2-enone O-methyl oxime ([18F]-FPECMO) is a novel derivative of[11C]-ABP688. [18F]-FPECMO was characterized as a PET imaging agent for the metabotropic glutamate receptor subtype 5 (mGluR5).Methods: [18F]-FPECMO was synthesized in a one-step reaction sequence by reacting [18F]-KF-K222 complex with (E)-3-((6-bromopyridin-2-yl)ethynyl)cyclohex-2-enone O-methyl oxime in dry DMSO. The in vitro affinity of FPECMO was determined by displacement assaysusing rat whole brain homogenates (without cerebellum) and the mGluR5-specific radioligand [3H]-M-MPEP. Further in vitrocharacterization involved metabolite studies, lipophilicity determination and autoradiographical analyses of brain slices. In vivo evaluationwas performed by postmortem biodistribution studies and PET experiments using Sprague-Dawley rats.Results: The radiochemical yield after semipreparative HPLC was 35±7% and specific activity was N240 GBq/μmol. [18F]-FPECMOexhibited optimal lipophilicity (logD=2.1) and high metabolic stability in vitro. Displacement studies revealed a Ki value of 3.6±0.7 nM forFPECMO. Biodistribution studies and ex vivo autoradiography showed highest radioactivity accumulation in mGluR5-rich brain regionssuch as the striatum and hippocampus. Co-injection of [18F]-FPECMO and ABP688 (1 mg/kg body weight), an mGluR5 antagonist, showed40% specific binding in the striatum, hippocampus and cortex, regions known to contain high densities of the mGluR5. PET imaging,however, did not allow the visualization of mGluR5-rich brain regions in the rat brain due to a fast washout of [18F]-FPECMO from mGluR5-expressing tissues and rapid defluorination.Conclusions: [18F]-FPECMO showed significant potential for the detection of mGluR5 in vitro; however, its in vivo characteristics are notoptimal for a clear-cut visualization of the mGluR5 in rats.© 2009 Elsevier Inc. All rights reserved.

Keywords: Positron emission tomography; Metabotropic glutamate receptor subtype 5; FPECMO; Radiosynthesis

1. Introduction

Glutamate is considered as the major excitatory neuro-transmitter in the central nervous system (CNS). Thereceptors activated by glutamate belong to a large familywhich can be divided into two groups: the ionotropicglutamate receptors and the metabotropic glutamate recep-tors. The ionotropic glutamate receptors (iGluRs) areglutamate-gated cation channels known as NMDA, kainate

⁎ Corresponding author. Tel.: +41 446337463; fax: +41 44 633 1367.E-mail address: simon.ametamey@pharma.ethz.ch (S.M. Ametamey).

0969-8051/$ – see front matter © 2009 Elsevier Inc. All rights reserved.doi:10.1016/j.nucmedbio.2009.03.005

or AMPA receptors depending on their agonist [1]. Themetabotropic glutamate receptor (mGluR) family compriseseight G-protein-coupled receptors named mGluR1 tomGluR8. These eight receptor subtypes have been classifiedinto three classes (I–III), based on their pharmacology,amino acid sequence and second messenger [2]. Group Iincludes mGluR1 and mGluR5; Group II includes mGluR2and mGluR3; Group III includes mGluR4, mGluR6,mGluR7 and mGluR8. The mGluR5 subtype, predominantlylocated in the hippocampus, striatum and cortex [3,4], issupposed to be implicated in CNS disorders such asdepression [5], anxiety [6–8] and Parkinson's disease [9,10].

Fig. 1. Chemical structures of FPECMO and some mGluR5 PET radioligands.

614 C. Lucatelli et al. / Nuclear Medicine and Biology 36 (2009) 613–622

Positron emission tomography (PET) is a noninvasiveimaging technology, which allows visualization and analysisof brain receptors using the appropriate PET radioligands.Our group recently described the first successful mGluR5PET ligand in rodents and human, [11C]-ABP688 1 [11](Fig. 1). This radiotracer, labeled with carbon-11, exhibitedstrong and specific mGluR5 signals in rodent brain andaccumulation in mGluR5-rich regions in human brain [12].However, for practical reasons, a fluorine-18 radiolabeledPET ligand is of particular interest. Several efforts towardsthe preparation of fluorine-18-labeled radiotracers have beenreported in the literature. Recently, Hamill et al. described[18F]-F-MTEB 3 [13] and [18F]-F-PEB 4 [13,14] (Fig. 1).Both ligands exhibited high binding affinities (Ki=80 pM and0.2 nM, respectively) and good in vivo properties. However,the radiochemical yields of both radiotracers were low (2–5%, decay corrected). Wang et al. [15] tried to improve theradiochemical yield of 4 by using thermal heating instead ofthe microwave approach used by Hamill et al. [13,14] butonly reached 5% radiochemical yield. Siméon et al. [16]described a fluoromethyl thiazol analog of F-MTEB 5, a newtracer offering a higher radiochemical yield. This mGluR5radioligand demonstrated a very high affinity (36 pM) andexcellent radiochemical yield (87%); however, quantificationof its uptake in monkey brain was limited. During thepreparation of this manuscript, Brown et al. [17] reported onthe evaluation of this radioligand in the human brain.

We recently published a new fluorine‑18 analog, [18F]-FE-DABP688 2 [18]. Unfortunately, this compound dis-played unfavorable pharmacokinetics resulting in a fastwashout from the brain and a short-lasting signal in the targetregions. The present work describes the synthesis, fluorine-18 labeling and the pharmacological evaluation of a 2-fluoropyridine analog of ABP688, [18F]-(E)-3-((6-fluoropyr-idin-2-yl)ethynyl)cyclohex-2-enone O-methyl oxime ([18F]-FPECMO) (6) (Fig. 1), as a potential mGluR 5 imaging agent.

2. Materials and methods

2.1. General

Solvents were purchased from Merck and Fluka and wereused without further purification. Chemicals were obtainedfrom Aldrich and Fluka. [3H]-MPEP was kindly providedby Novartis.

The semipreparative HPLC system used for the tracerpurification consisted of a Knauer pump, a Knauerultraviolet detector and a Geiger Müller LND 714 counterwith an Eberlein RM-14 instrument as well as a μBondapak(250×10 mm, Waters) column using isocratic elution (flow, 4ml/min) and a mobile phase consisting of 35% acetonitrileand 65% water containing 0.1% H3PO4.

Quality control of the final product was performed on anAgilent system equipped with a radiodetector from Raytest

615C. Lucatelli et al. / Nuclear Medicine and Biology 36 (2009) 613–622

using a Bondclone 10 μm C18 (300×3.9 mm; Phenomenex)column and applying a linear gradient of HPLC eluents Aand B. Eluent A consisted of water containing 0.1% oftrifluoroacetic acid; eluent B consisted of acetonitrilecontaining 0.1% of trifluoroacetic acid. The HPLC gradientwas 20% to 80% of eluent B over 20 min, at a flow rate of1.5 ml/min.

Metabolites analyses were performed on a WatersAcquity UPLC system equipped with a radiodetectorfrom Berthold and an Acquity UPLC BEH C18 1.7 μm(2.1×50 mm) column. An isocratic flow consisting of amixture of acetonitrile (40%), water (60%) and TFA (0.1%)at 0.914 ml/min was applied.

Animal care and all experimental procedures wereapproved by the Swiss Federal Veterinary Office. Animals(male Sprague-Dawley rats, 250–450 g, obtained fromCharles River) were allowed free access to food and water.

2.2. Chemistry

2.2.1. 3-Ethynylcyclohex-2-enone O-methyl oxime (9 )To a solution of 8 (427 mg, 3.55 mmol) in pyridine (10

ml) at room temperature (RT) was added O-methylhydroxylamine hydrochloride (445 mg, 5.32 mmol). Theabove solution was stirred at RT for 18 h before water (30ml) was added. The reaction mixture was then extracted withEt2O (2×20 ml) and the combined organic layers werewashed with sat. CuSO4 solution (4×20 ml). The organiclayer was washed with brine, dried over Na2SO4 andevaporated under reduced pressure. Purification of the crudemixture by column chromatography (Et2O/pentane 1:9 v/v)afforded the E- and Z-isomers of 9 in 42% (226 mg) and16% (84 mg) yield, respectively.

1H NMR (400 MHz, CDCl3, E-isomer) δ 6.43 (t, J=1.36Hz, 1H), 3.87 (s, 3H), 3.13 (s, 1H), 2.5–2.46 (m, 2H), 2.27–2.24 (m, 2H), 1.76–1.70 (m, 2H).

1H NMR (400 MHz, CDCl3, Z-isomer) δ 7.02 (t, J=1.72Hz, 1H), 3.88 (s, 3H), 3.17 (s, 1H), 2.35–2.29 (m; 4H),1.85–1.79 (m, 2H).

13C-NMR (400 MHz, CDCl3, E-isomer) δ 155.2, 130.8,127.1, 84.4, 80.5, 61.9, 29.4, 22.0, 20.7.

2.2.2. (E)-3-((6-Bromopyridin-2-yl)ethynyl)cyclohex-2-en-one O-methyl oxime (10 )

To a carefully degassed solution of 2,6-dibromopyridine(730 mg, 3.08 mmol), E-isomer 9 (230 mg, 1.54 mmol), CuI(52 mg, 0.27 mmol) in Et3N (12 ml) and DMF (1 ml) wasadded Pd(PPh3)4 (103 mg, 0.09 mmol). The reaction mixturewas stirred at RT for 12 h before saturated aqueous NH4Clsolution (10 ml) was added. The mixture was extracted withEtOAc (3×20 ml) and the organic phase was thensuccessively washed with water (4×20 ml), brine (2×20ml) and dried over Na2SO4. Evaporation of the solventsunder reduced pressure afforded the crude product whichwas purified by flash chromatography [flash silica gel,pentane-EtOAc (80/20)] to yield 10 (324 mg, 69%) as abrown solid.

Mp.: 83.3°C1H-NMR (400 MHz, CDCl3) δ 1.76–1.82 (m, 2H), 2.35–

2.39 (m, 2H), 2.51–2.55 (m, 2H), 3.92 (s, 1H), 6.57(t, J=1.58 Hz, 1H), 7.37–7.43 (m, 2H), 7.48–7.52 (m, 1H).

13C-NMR (100 MHz, CDCl3) δ 21.1, 21.4, 29.5, 62.4,90.8, 91.8, 126.4, 127.0, 127.8, 132.0, 138.6, 142.2, 144.0,155.5.

MS (ESI) m/z 308 (15%), 307 (MH+, 100%), 306 (15%),305 (100%).

2.2.3. (E)-3-((6-Fluoropyridin-2-yl)ethynyl)cyclohex-2-en-one O-methyl oxime (6 )

To a carefully degassed solution of the E-isomer of 9(55.4 mg, 0.37 mmol) and 2-bromo-6-fluropyridine (150 mg,0.85 mmol) in Et3N (10 ml) and DMF (1 ml) were added Pd(PPh3)4 (25.4 mg, 0.02 mmol) and CuI (17.7 mg, 0.07mmol). The reaction mixture was stirred at RT under anargon atmosphere for 20 h before saturated aqueous NH4Clsolution (10 ml) was added. The above mixture wasextracted with EtOAc (3×20 ml) and the organic layer waswashed with water (4×20 ml) and brine (25 ml) before it wasdried over Na2SO4. Evaporation of the solvents underreduced pressure gave the crude product which was purifiedby flash chromatography (flash silica gel, pentane-EtOAc(80/20)) to yield 6 (28 mg, 31%) as an oil.

1H NMR: (400 MHz, CDCl3) δ 1.77 (quint., J=6.26 Hz,2H), 2.35 (dt, J=1.81 Hz, J=6.24, 2H), 2.51 (t, J=6.76 Hz,2H), 3.89 (s, 3H), 6.55 (t, J=1.52 Hz, 1H), 6.86 (dd, J=2.72Hz, J=8.13 Hz), 7.30 (dd, J=2.26 Hz, J=7.22 Hz, 1H), 7.72(q, J=8.4 Hz, 1H).

13C NMR (100 MHz, CDCl3) δ 20.7, 22.1, 29.2, 62.1,90.5, 91.0, 109.2, 109.6, 124.8, 126.7, 131.6, 141.2, 155.3,161.8.

19F-NMR (376 MHz, CDCl3) δ −65.35 (s, 1F).MS (ESI) m/z 245 (100%, MH+).

2.3. Radiochemistry

2.3.1. Synthesis of [18F]-FPECMO (6 )18F− was obtained via the 18O(p,n)18F reaction on 98%

enriched 18O-water. Aqueous 18F− from cyclotron wasinitially trapped on a light QMA cartridge (Waters)preconditioned with 0.5 M K2CO3 (5 ml) and water (5 ml).18F− was eluted from the cartridge into a 10-ml tightly closedreactive vial with a solution prepared from Kryptofix K222

(15 mg; Merck) and K2CO3 (3 mg) in 1.5 ml of acetonitrile/water (8:2). The solvents were evaporated at 110°C undervacuum and a stream of nitrogen. Dry acetonitrile (1 ml) wasthen added to the reactive vial and the whole mixture wasazeotropically evaporated to dryness. A solution of (E)-3-((6-bromopyridin-2-yl)ethynyl)cyclohex-2-enone O-methyloxime 10 (6 mg, 0.01 mmol) in dry DMSO (0.5 ml) wasadded to the dry K222-[

18F]KF complex and the reactionmixture was heated at 130°C for 10 min before water (2 ml)was added. Purification of the product was achieved onsemipreparative HPLC using conditions described underSection 2.1. The fraction containing the product was

616 C. Lucatelli et al. / Nuclear Medicine and Biology 36 (2009) 613–622

collected on a C18 Plus SepPak cartridge (Waters) and thecartridge was then washed with 20 ml of water in order toremove traces of acetonitrile. [18F]-FPECMO was elutedfrom the cartridge with 2 ml of ethanol and buffered with 8ml 0.15 M phosphate buffer to yield an injectable solutionafter sterile filtration.

2.4. Determination of distribution coefficient (logD)

The lipophilicity of [18F]-FPECMO was determined atphysiological pH using the shake-flask method as describedby Strijckmans et al. [19]. Briefly, [18F]-FPECMO (1.5 MBqin 100 μl) was mixed with 900 μl of octanol and 1 ml ofphosphate buffer (0.15 mM, pH 7.4). The radioactivityincorporated in both phases was determined using a gammacounter (Wizard, Perkin Elmer).

2.5. In vitro metabolite analysis

In vitro plasma stability was determined by incubating theradioligand in rat or human plasma at 37°C for 45 min.Proteins were precipitated using acetonitrile, and the super-natant was analyzed by analytical HPLC, applying condi-tions described previously.

The metabolic stability of the radiotracer was alsoassessed in liver microsomes by using pooled livermicrosomes from human or rat origin (BD Bioscience).The assay was carried out in 0.1 M phosphate buffer (pH 7.4)at 37°C, in the presence of 10 mM NADPH (replaced byPBS for negative control) and with the test compound (20–50 nM) or 15 μM testosterone as a positive control. After 5min of preincubation, the reaction was started by adding themicrosomal suspension (0.52 mg/ml protein final) atdifferent time points (0, 5, 20 and 60 min) at 37°C. Thereaction was stopped by adding 150 μl ice-cold acetonitrile.Samples were then centrifuged at 10,000×g for 5 min at 4°C,and the supernatant was analyzed by analytical HPLC andapplying conditions described above.

2.6. In vitro binding assays

2.6.1. Preparation of membranesMale rats were euthanized by decapitation and the brains

were quickly removed individually. Whole brains withoutcerebellum were homogenized in 10 volumes of ice-coldsucrose buffer (0.32 mol/L sucrose; 10 mmol/L Tris/acetatebuffer pH 7.4) with a PT 1200 C Polytron (Kinematica AG)for 1 min at setting 4. The homogenate was centrifuged at1000×g for 15 min (4°C) to yield a crude pellet (P1). Thispellet was resuspended in 5 volumes of sucrose buffer,homogenized and centrifuged again at 1000×g for 15 min(4°C). The resulting supernatants were combined andcentrifuged at 17,000×g for 20 min (4°C) to yield pelletP2. P2 was washed with ice-cold incubation buffer 1(5 mmol/L Tris/acetate buffer, pH 7.4), homogenized andcentrifuged at 17,000×g for 20 min (4°C). The P2 membranepellet was resuspended in incubation buffer 1 and stored at−70°C. On the day of the essay, the P2 membranes were

thawed and the protein concentration determined by a Bio-Rad micro-assay with bovine serum albumin as a standard.

2.6.2. Displacement assaysThe P2 membranes (100 μg/tube) were incubated with

2 nM [3H]-M-MPEP (2.3 GBq/μmol) and increasingconcentrations of FPECMO (1 pM–100 μM) in incubationbuffer 2 (30 mM NaHEPES, 110 mM NaCl, 5 mM KCl, 2.5mM CaCl2xH2O, 1.2 mM MgCl2, pH 8) to give a totalvolume of 200 μl. Nonspecific binding was determined in thepresence of 100 μM M-MPEP. Incubations were allowed toproceed for 45 min at RT before being terminated by vacuumfiltration over GF/C filters (Whatman). The filters were rinsedtwice with 4 ml of ice-cold incubation buffer 2 before 4 ml ofscintillation liquid (Ultima Gold, Perkin Elmer) was added.The radioactivity retained on the filters was determined usinga beta counter (Liquid Scintillation Analyser 1900TR,Canberra Packard). Binding data were evaluated using theKELL Radlig Software (Biosoft). Three independent experi-ments were performed, and the inhibition constant Ki wascalculated using the Cheng–Prusoff equation.

2.7. In vitro autoradiography

The brains of Sprague-Dawley rats and 129 mice wereremoved and frozen in 2-methylbutane (Fluka) at −30 to−36°C. The frozen brains were cut at−18°C into 20-μm-thickhorizontal slices using the Cryostat microtome HM 505N(Microm) and put on SuperFrost slides (Menzel). The slideswere stored at −80°C until use. On the day of the experiment,the sections were thawed for 30 min at RTand then incubatedin HEPES-BSA buffer (30 mM Na-HEPES, 110 mM NaCl,5 mM KCl, 2.5 mM CaCl2xH2O, 1.2 mMMgCl2, 0.1%BSA,pH 7.4) for 10min on ice. The slides were then incubated witha 0.4-nM [18F]-FPECMO solution for 45 min at RT. Theslides were washed three times (3 min each) with HEPESbuffer and twice with distilled water (5 s each). The slideswere then dried for 20 min at RT and then exposed for 30–45min to the phosphor imager plates and the plates weresubsequently scanned by the BAS5000 reader (Fuji).

2.8. In vivo PET imaging

PET imaging of three Sprague-Dawley rats (199–327 g)was performed using the 16-module variant of the quad-HIDAC tomograph (Oxford Positron Systems, UK). Beforeinjection of the radioligand, the animals were anesthetizedwith isoflurane and placed and fixed in the field of view ofthe tomograph. Monitoring of anesthesia was performedaccording to a literature procedure [20]. [18F]-FPECMO(16–18 MBq; b0.15 nmol) was administered via a tail veincatheter, and PET data acquisition was initiated simulta-neously. Scan duration was 90 min for two experiments and20 min for one experiment. PET data were acquired in listmode and reconstructed in user-defined time frames (5×2,6×5, 5×10 min). The bin size was 1.0 mm with a matrix sizeof 250×100×100 for whole body images. Image files were

cheme 1. Synthesis of FPECMO. Reagents and conditions: (a) ethynyl-agnesium bromide, THF, RT; (b) methoxylamine hydrochloride, pyridine,T; (c) 2,6-dibromopyridine or 2-bromo-6-fluoropyridine, Pd(Ph3)4, CuI,t3N, DMF, RT.

617C. Lucatelli et al. / Nuclear Medicine and Biology 36 (2009) 613–622

evaluated by region-of-interest (ROI) analysis using thededicated software PMOD [21]. Time–activity curves werenormalized to the injected dose per gram of body weight andexpressed as standardized uptake values.

2.9. Postmortem biodistribution studies

[18F]-FPECMO was administered to restrained Sprague-Dawley rats (192–250 g) via a lateral tail vein. Animalswere injected with a constant volume (300 μl) of [18F]-FPECMO corresponding to an injected tracer mass ofapproximately 80 pmol. The animals were sacrificed bydecapitation at 5 and 30 min after injection. Blockadestudies were carried out by co-injecting ABP688 (1.0 mg/kg body weight; 1.5 mg/ml PEG/H2O=50/50) into threeanimals. The control group (n=3) was injected with acorresponding volume of vehicle (PEG/H2O=50/50).Whole brains were rapidly removed and dissected intospecific brain regions: hippocampus, striatum, cortex,cerebellum, midbrain and the rest of the brain. Blood,urine, faeces, spleen, lung, liver, kidney, muscle, heart,adipose tissue, harderian glands, stomach, intestine,pancreas and bone were also taken. Individual brainregions and organ samples were weighed and tissueradioactivity was measured in a gamma-counter (Wizard,Perkin Elmer). For all samples taken, the activityconcentration was determined using the percent injecteddose per gram wet tissue (%ID/g tissue).

2.10. In vivo metabolism

[18F]-FPECMO was injected into restrained rats via alateral tail vein (500 μl, 39–140 MBq). The animals weresacrificed by decapitation at 5 or 30 min postinjection.Whole brains were rapidly harvested and cerebella werediscarded. Brains were then homogenized for 1 min with apolytron in 1 ml of phosphate buffer solution (0.15 M) beforeacetonitrile (1 ml) was added. Brain homogenate suspen-sions were agitated with vortex for 1 min followed bycentrifugation at 4800 rpm for 5 min. Supernatants werecollected and analyzed by UPLC using the previouslydescribed method. Blood was collected in heparin-coatedtube and plasma was separated by centrifugation at 4800 rpmfor 5 min. The supernatant (750 μl) was collected andproteins were precipitated by addition of acetonitrile (750 μl)before centrifugation for 5 min at 4800 rpm. The supernatantwas collected and analyzed by UPLC. Urine was sampled bypunctuation of the bladder. Acetonitrile (750 μl) was addedand the mixture was centrifuged at 4800 rpm for 5 min. Thesupernatant was collected and analyzed by UPLC.

2.11. Ex vivo autoradiography

A Sprague-Dawley rat was injected with 48 MBq [18F]-FPECMO and sacrificed by decapitation 5 min later. Thebrain was removed and frozen in 2-methylbutane (Fluka) at−30°C to −36°C. The frozen brain was immediately cut at−18°C into 20-μm-thick horizontal slices using the Cryostat

microtome HM 505N (Microm) and put on SuperFrost slides(Menzel). Slices were then exposed for 30 min to thephosphor imager plates and the plates were subsequentlyscanned by the BAS5000 reader (Fuji).

3. Results

3.1. Chemistry

The syntheses of precursor 10 and reference compound 6were accomplished using the same three-step convergentroute (Scheme 1). The first step of the synthesis involvingthe preparation of 3-ethynylcyclohex-2-enone (8) wasperformed essentially in analogy to the method describedby Larsen and O'Shea [22]. Compound 8was then convertedinto methyl oxime 9 by reaction with O-methyl hydro-xylamine. The separation of both the Z and E isomers waseasily achieved using flash chromatography and the yield forthe thermodynamically favored E isomer was greater thanfor the Z isomer (42% vs. 16%). The E-isomer was identifiedby NMR in analogy to cyclohexenone methyl oxime [23]. Inthe 1H NMR spectra of the oxime ether, the signal of thevinyl proton in the α-position to the oxime functionalityappears at a lower field for the Z-isomer. The Sonogashiracross-coupling reaction [24] of the alkyne 9 with 2-bromo-6-fluoropyridine or 2,5-dibromopyridine afforded 6 and thebrominated precursor 10, respectively. After this couplingreaction, no isomerization at the oxime functionality wasobserved by NMR.

3.2. In vitro binding assay

To determine the affinity of FPECMO for mGluR5,displacement assays with the established mGluR5 radioli-gand [3H]-M-MPEP [25] were realized using rat brainmembranes. Nonspecific binding was quantified in thepresence of 100 μM M-MPEP. FPECMO exhibited a high

SmRE

Fig. 2. Displacement of [3H]-M-MPEP by [19F]-FPECMO in rat whole brain membranes (without cerebellum). Values are presented as percentage of specibinding and represent the mean±S.D. of three independent experiments performed in triplicates.

618 C. Lucatelli et al. / Nuclear Medicine and Biology 36 (2009) 613–622

affinity for mGluR5 (Ki=3.6±0.7 nM) as shown in Fig. 2.This affinity is comparable to the benchmark compoundABP688 which was determined in parallel assays (Ki=3.1±0.8 nM). These results demonstrate that the introduction ofa fluorine atom on the pyridine ring does not compromise theaffinity. This promising in vitro property of FPECMOprompted us to radiolabel this compound with fluorine-18for further studies.

3.3. Radiosynthesis of [18F]-FPECMO

[18F]-FPECMO was radiolabeled with fluorine-18 in aone‑step procedure in analogy to previously describedmethods [26] (Scheme 2). [18F]‑FPECMO was obtained ingreater than 99% radiochemical purity, and the specificradioactivity was greater than 240 GBq/μmol. The totalsynthesis time was 50 min, and the radiochemical yield was35% (nN10). The identity of [18F]-FPECMO was confirmedby co-injection with the reference FPECMO which gaveexactly the same retention (5.6 min) time under identicalelution conditions.

3.4. Determination of distribution coefficient (logD)

The lipophilicity of [18F]-FPECMO was determinedexperimentally at physiological pH using the shake-flask

Scheme 2. Radiosynthesis of [18F]-FPECMO. Reagents and conditions: (a) [18F]KF-K222, DMSO, 130°C.

fic

method. The obtained logD value (2.11±0.13) was compar-able to the calculated logP value (cLogP=2.16).

3.5. Metabolic stability in vitro

No radioactive degradation products of [18F]-FPECMOwere observed after 90 min of incubation at 37°C in plasmaor whole blood from human or rat. Likewise, the stabilitystudies of [18F]-FPECMO in human or rat liver microsomesdid not reveal any radioactive metabolites.

3.6. In vitro autoradiography

In vitro autoradiographical studies realized with murineand rat brain slices demonstrated the specificity of [18F]-FPECMO binding to mGluR5. Heterogeneous uptake ofthe radiotracer in mGluR5-rich regions such as thehippocampus, striatum and cortex (Fig. 3) was observed.As expected, radioactivity accumulation in the cerebellumwas negligible which confirms the low mGluR5 expressionin this region. In addition, blocking experiments usingunlabeled ABP688 and [18F]-FPECMO confirmed thespecificity of [18F]-FPECMO for mGluR5 whereby amarkedly reduced accumulation of [18F]-FPECMO wasobserved in both mouse and rat brain slices.

Fig. 3. In vitro [18F]-FPECMO autoradiography of horizontal sections ofmouse and rat brains under baseline and blocking conditions (co-incubationwith 10 μM ABP688).

Fig. 4. Biodistribution of [18F]-FPECMO in rats under baseline and blockade condit(%ID/g).

619C. Lucatelli et al. / Nuclear Medicine and Biology 36 (2009) 613–622

3.7. In vivo characterization

3.7.1. Biodistribution studiesPostmortem biodistribution studies under both baseline

and blockade conditions were undertaken to demonstrate thespecificity of [18F]-FPECMO in vivo. Rats were injectedwith the radiotracer and sacrificed at 5 and 30 minpostinjection. For both time points, analyzed highest activityaccumulation was identified in mGluR5-rich brain regionssuch as the striatum, hippocampus and cortex, whereasradioactivity uptake in the cerebellum was clearly lower(Fig. 4). The mean striatum-to-cerebellum uptake ratios were2.39±0.20 and 2.32±0.09 at 5 and 30 min postinjection,respectively. Specificity of binding was also confirmed byblockade experiments with ABP688, an antagonist formGluR5. Up to 40% specific binding was observed in thestriatum and hippocampus at both time points analyzed. Asexpected, no blocking effects were evident in the cerebellumand midbrain regions. Brain activity concentrations clearlydecreased between 5 and 30 min postinjection (e.g., from0.90 %ID/g to 0.18 %ID/g for the striatum in controlanimals) pointing to a rapid clearance of the radiotracer fromthe brain. In contrast, target-to-nontarget ratios as well as thedegree of specific binding did not largely change between 5and 30 min after tracer administration.

No significant blocking effect was observed for allperipheral organs tested (Table 1). Highest activity concen-trations at 5 min postinjection were identified in excretoryorgans such as kidney and liver as well as in the brain-neighboring Harderian gland. At 30 min postinjection,highest activity concentrations were identified in urine,faeces and bone. These data suggest that [18F]-FPECMO is

ions at 5 and 30 min pi. Data are expressed as % injected dose per gram tissue

ig. 6. (A) Time–activity curves for bone, cortex/striatum/hippocampus anderebellum of a rat injected with 17.9 MBq [18F]-FPECMO and scanned for0 min. (B) Detail of (A) highlighting the time–activity curves for cortex/triatum/hippocampus and cerebellum.

Table 1Biodistribution of [18F]-FPECMO in rats under baseline and blockadeconditions at 5 and 30 min pi

Control(5 min pi)

Blockade(5 min pi)

Control(30 min pi)

Blockade(30 min pi)

Spleen 0.33±0.04 0.26±0.01 0.19±0.06 0.16±0.02Liver 1.08±0.43 1.35±0.53 0.63±0.03 0.65±0.10Kidneys 1.37±0.52 1.11±0.40 0.58±0.05 0.63±0.17Lung 0.60±0.18 0.47±0.14 0.30±0.06 0.26±0.03Bone 0.37±0.06 0.24±0.04 1.39±0.09 1.01±0.15Heart 0.44±0.08 0.40±0.05 0.15±0.01 0.15±0.02Adipose tissue 0.24±0.03 0.20±0.07 0.37±0.09 0.48±0.08Harderian gland 1.15±0.18 1.41±0.21 0.76±0.12 0.89±0.14Muscle 0.35±0.03 0.30±0.02 0.14±0.01 0.13±0.02Blood 0.39±0.10 0.31±0.05 0.18±0.01 0.17±0.01Stomach 0.31±0.08 0.28±0.05 0.15±0.01 0.15±0.02Intestine 0.41±0.10 0.42±0.07 0.27±0.03 0.29±0.13Pancreas 0.59±0.19 0.56±0.05 0.22±0.04 0 28±0 08Urine 0.17±0.12 0.22±0.12 0.64±0.56 0.93±0.80Faeces 0.17±0.10 0.84±0.57 2.53±1.86 0.97±0.64

Data are expressed as % injected dose per gram of tissue.

620 C. Lucatelli et al. / Nuclear Medicine and Biology 36 (2009) 613–622

metabolized via renal and hepatobiliary pathways and thatbone-seeking [18F]-fluoride is a probable radiometabolite.

3.7.2. PET StudiesA dynamic PET study was performed to analyze whether

mGluR5-expressing regions in the rat brain can be visualizedby [18F]-FPECMO. Data were acquired from 0 to 90 minpostinjection and summed images for a single 90-min timeframe revealed very low brain uptake and considerably higheractivity accumulation in bone and joints. Even for early timeframes immediately after tracer injection, radioactivityaccumulation in the brain was insufficient to yield a clearsignal (Fig. 5). ROI analysis of dynamic PET data confirmed

Fig. 5. Comparison of the image quality of [18F]-FPECMO PET dataacquired from 0 to 10 min pi (left) and a corresponding midframe ex vivoautoradiographical image (right).

Fc9s

that radioactivity uptake in the skeleton strongly increasedduring the duration of the scan (Fig. 6A). In contrast, activityin forebrain (cortex, striatum, hippocampus) and hindbrain(cerebellum) regions peaked immediately after radiotracerinjection followed by a rapid washout (Fig. 6B).

3.7.3. In vivo metabolismAnimals were sacrificed 5 and 30 min after radiotracer

injection, and brain, blood as well as urine samples werecollected and prepared for UPLC analysis. In brain samples,only a single radioactivity peak corresponding to nonmeta-bolized [18F]-FPECMO was evident at both time pointsanalyzed. Fig. 7 shows a representative UPLC chromatogramof a brain sample collected at 5 min postinjection. Plasmasamples showed a second radioactivity peak besides the peakof the parent compound. The polar radioactive metabolitemost probably corresponded to [18F]-fluoride and increasedbetween 5 and 30 min postinjection. In urine samples, onlythe polar radioactive metabolite was identified.

4. Discussion

Up- or down-regulation of mGluR5may be implicated in avariety of psychiatric and neurodegenerative diseases [5–10].

Fig. 7. A typical UPLC chromatogram of a brain sample 5 min after injection.

621C. Lucatelli et al. / Nuclear Medicine and Biology 36 (2009) 613–622

A noninvasive method such as PET should be ideally suitedto analyzing and identifying these changes in mGluR5density which might then allow the diagnosis of thesedisorders in a presymptomatic state. The short half-life ofcarbon-11 (20.4 min) precludes the transport of the radio-tracer from its production site to other PET imaging siteslacking cyclotron facilities. Thus, the development of afluorine-18 radiolabeled tracer for mGluR5 is of high interest.

The synthesis of FPECMO started with the keyintermediate 8 which was obtained from the reaction ofethynylmagnesium bromide with 3-ethoxy-2-cyclohexen-1-one (7). (E)-3-Ethynylcyclohex-2-enone O-methyl oxime(9) was prepared in 42% yield from 3-ethynylcyclohex-2-enone (8) after a simple condensation reaction with O-methyl hydroxylamine (Scheme 1). The E and Z isomerswere easily separated by flash chromatography, and theassignment of the isomers was based on NMR chemicalshifts and also by comparison with existing structures in theliterature [23]. Separation of the E and Z isomers wasnecessary, since the E isomer of ABP688 was shown inprevious studies to be more potent than the Z-isomer [11].Although this may not necessarily be true for thecompounds currently under study, we nevertheless carriedout the next steps of the synthesis with only the E-isomer.Indeed, the NMR chemical shift of the vinyl proton wasconsiderably modified by the relative cis or trans positionof the methoxy group so that the signal of the vinyl protonof the Z isomer was shifted to a lower field (chemical shifthigher in parts per million) when compared to the signal ofthe E isomer. Target compound 6 and the bromo-precursor10 were obtained in 31% and 69% yield, respectively, afterreacting intermediate 9 with either 2-bromo-6-fluoropyr-idine or 2,6-dibromopyridine using a Sonogashira cross-coupling reaction.

The radiosynthesis of [18F]-FPECMO was achieved in aone-step reaction sequence by reacting [18F]KF-K222

complex with the bromo precursor 10. Pure radiolabeledtracer was obtained after semipreparative HPLC purifica-tion. In all preparations, the final compound was obtainedin good radiochemical yield (35±7% decay corrected,nN10). The identity of the final compound was confirmedby co-elution with the nonradioactive reference compoundusing analytical HPLC.

The lipophilicity of the tracer was determined experi-mentally using the shake-flask method. The measured logDvalue of 2.1 suggests that [18F]-FPECMO is sufficientlylipophilic for free diffusion through the blood–brain barrier(BBB). For CNS PET ligands, a postulated logD valuebetween 2 and 3 has been given as an optimal range for goodBBB penetration [27].

FPECMO showed high affinity binding for the mGluR5 invitro in rat brain membranes (Ki=3.6±0.7 nM) and comparesfavourably with the reported value of ABP688 (Ki=3.1±0.8nM) [11]. From these data, we concluded that the in vitroaffinity of FPECMO was not significantly different from thatof ABP688 and that further investigation was warranted.

Initially, we performed in vitro autoradiographical studieswith [18F]-FPECMO using both murine and rat brain slicesand observed a heterogeneous radioactivity accumulation indifferent regions of the brain. As expected, highest traceruptake was localized in mGluR5-rich regions such as thehippocampus, striatum and cortex. Blockade studies usingABP688 demonstrated the specificity of the novel tracer formGluR5 and a markedly reduced accumulation of [18F]-FPECMO was observed in both mouse and rat brain slices(Fig. 3).

Biodistribution studies and ex vivo autoradiographiesconfirmed this specificity revealing highest radioactivityaccumulation in mGluR5-rich brain regions. A 40% blockingeffect was observed when the radiotracer was co-injected withABP688 (1 mg/kg body weight). Despite these favorable invitro and ex vivo characteristics, [18F]-FPECMOdid not allowthe visualization of mGluR5-rich brain regions in the rat brainin vivo. A kinetic analysis of PET data showed a very fastwashout from the mGluR5-expressing brain regions. Even forearly time points immediately after tracer injection (e.g., for a0- to 10-min time frame), radioactivity accumulation in thebrainwas insufficient for a clear PETsignalwith an appropriatesignal-to-noise ratio (Fig. 5). In contrast, ex vivo autoradio-graphy performed at 5 min after injection demonstrated aspecific signal in the striatum and hippocampus suggesting thatthe sensitivity of the PET system compared to the autoradio-graphical imager plates may be inadequate to produce high-quality images with clear-cut signals. In addition, ex vivometabolite analyses taken at 5min postinjection confirmed thatradioactivity uptake in the brain was related to nonmetabolized

622 C. Lucatelli et al. / Nuclear Medicine and Biology 36 (2009) 613–622

[18F]-FPECMO (Fig. 7) and that almost the entire radioactivityfound in the plasma at 5 min pi was still attributable to thenonmetabolized parent compound. Therefore, the metabolicdegradation (i.e., defluorination) of [18F]-FPECMO is not theonly factor contributing to the poor in vivo characteristics ofthis PET tracer. Insufficient penetration of the BBB, strongefflux from the brain by a PgP-transport mechanism as well asunfavorable in vivo kinetic parameters of the ligand–receptorinteraction (i.e., kon, koff) might jointly contribute to aninadequate accumulation of the radiotracer in mGluR5-expressing brain regions.

In conclusion, [18F]-FPECMO is a novel radioligand formGluR5. The easy synthesis of the precursor as well as theone-step radiolabeling procedure makes [18F]-FPECMO anattractive ligand for in vitro work on mGluR5. Despite highmetabolic stability in various in vitro test systems, a strongdefluorination of the radiolabeled substance was identified inrats in vivo, which challenges the significance and validity ofsuch in vitro tests during CNS PET tracer development.Since metabolic patterns are known to differ among species,a more favorable kinetic profile might be anticipated for thehuman situation.

Acknowledgment

The authors thank Marianne Kehl, Manuela Grueninger,Claudia Keller and Sabine Baumann for excellent experi-mental help.

References

[1] Dingledine R, Borges K, Bowie D, Traynelis SF. The glutamatereceptor ion channels. Pharmacol Rev 1999;51:7–61.

[2] Pin JP, Duvoisin R. The metabotropic glutamate receptors: structureand functions. Neuropharmacology 1995;34:1–26.

[3] Abe T, Sugihara H, Nawa H, Shigemoto R, Mizuno N, Nakanishi S.Molecular characterization of a novel metabotropic glutamate receptormGluR5coupled to inositol phosphate/Ca2+ signal transduction. J BiolChem 1992;267:13361–8.

[4] Shigemoto R, Nomura S, Ohishi H, Sugihara H, Nakanishi S, MizunoN. Immunohistochemical localization of a metabotropic glutamatereceptor, mGluR5, in the rat brain. Neurosci Lett 1993;163:53–7.

[5] Palucha A, Branski P, Szewczyk B, Wieronska JM, Klak K, Pilc A.Potential antidepressant-like effect of MTEP, a potent and highlyselective mGluR5 antagonist. Pharmacol Biochem Behav 2005;81:901–6.

[6] Allgeier H, Cosford ND, Flor PJ, et al. mGluR5 antagonists for thetreatment of pain and anxiety. European Patent 2003 [EP01 117 403].

[7] Spooren W, Gasparini F. mGlu5 receptor antagonists: a novel class ofanxiolytics. Drug News Perspect 2004;17:251–7.

[8] Tatarczynska E, Klodzinska A, Chojnacka-Wójcik E, et al. Potentialanxiolytic- and antidepressant-like effects of MPEP, a potent, selectiveand systemically active mGlu5 receptor antagonist. Br J Pharmacol2001;132:1423–30.

[9] Rouse ST, Marino MJ, Bradley SR, Awad H, Wittmann M, Conn PJ.Distribution and roles of metabotropic glutamate receptors in the basalganglia motor circuit: implications for treatment of Parkinson's diseaseand related disorders. Pharmacol Ther 2000;88:427–35.

[10] Marino MJ, Conn JP. Modulation of the basal ganglia by metabotropicglutamate receptors: potential for novel therapeutics. Curr DrugTargets CNS Neurol Disord 2002;1:239–50.

[11] Ametamey SM, Kessler LJ, Honer M, et al. Radiosynthesis andpreclinical evaluation of 11C-ABP688 as a probe for imaging themetabotropic glutamate receptor subtype 5. J Nucl Med 2006;47:698–705.

[12] Ametamey SM, Treyer V, Streffer J, et al. Human PET studies ofmetabotropic glutamate receptor subtype 5 with 11C-ABP688. J NuclMed 2007;48:247–52.

[13] Hamill TG, Krause S, Ryan C, et al. Synthesis, characterization, andfirst successful monkey imaging studies of metabotropic glutamatereceptor subtype 5. Synapse 2005;56:205–16.

[14] Patel S, Hamill TG, Connolly B, Jagoda E, Li Gibson RE. Speciesdifferences in mGluR5 binding sites in mammalian central nervoussystem determined using in vitro binding with [18F]F-PEB. Nuc MedBiol 2007;34:1009–17.

[15] Wang JQ, Tueckmantel W, Zhu A, Pellegrino D, Brownwell AL.Synthesis and preliminary biological evaluation of 3-[18F]fuoro-5-(2-pyridinylethynyl)benzonitrile as a PET radiotracer for imagingmetabotropic glutamate receptor subtype 5. Synapse 2007;61:951–61.

[16] Siméon FG, Brown AK, Zoghbi SS, Patterson VM, Innis RB, PikeVW. Synthesis and simple 18F-labeling of 3-fluoro-5-(2-(2-(fluoro-methyl)thiazol-4-yl)ethynyl)benzonitrile as a high affinity radioligandfor imaging monkey brain metabotropic glutamate subtype-5 receptorswith positron emission tomography. J Med Chem 2007;50:3256–66.

[17] Brown AK, Kimura Y, Zoghbi SS, Siméon FG, Liow JS, Kreisl WC,et al. Metabotropic glutamate subtype 5 receptors are quantified inthe human brain with a novel radioligand for PET. J Nucl Med 2008;49:2042–8.

[18] Honer M, Stoffel A, Kessler LJ, Schubiger PA, Ametamey SM.Radiolabeling and in vitro and in vivo evaluation of [18F]-FE-DABP688 as a PET radioligand for the metabotropic glutamatereceptor subtype 5. Nuc Med Biol 2007;34:973–80.

[19] Strijckmans V, Hunter DH, Dolle F, Coulon C, Loc'h C, Mazière B.Synthesis of a potential M1 muscarinic agent [76Br]bromocaramiphen.J Labelled Compd Radiopharm 1996;38:471–81.

[20] Honer M, Bruhlmeier M, Missimer J, Schubiger AP, Ametamey SM.Dynamic imaging of striatal D2 receptors in mice using quad-HIDACPET. J Nucl Med 2004;45:464–70.

[21] Mikolajczyk K, Szabatin M, Rudnicki P, Grodzki M, Burger C. AJAVA environment for medical image data analysis: initial applicationfor brain PET quantitation. Med Inform (Lond) 1998;23:207–14.

[22] Larsen DS, O'Shea MD. Synthetic approaches to the angucyclineantibiotics: the total syntheses to (±)-rubiginone B1 and B2, (±)-emycinA, and related analogs. J Chem Soc Perkin Trans 1 1995;8:1019–28.

[23] Zaidlewicz M, Uzarewicz IG. Reduction of O-methyl oxime ethers ofconjugated cyclohexenones with aluminium hydride. Heteroat Chem1993;4:73–7.

[24] Sonogashira K, Tohda Y, Hagihara NA. Convenient synthesis ofacetylenes: catalytic substitutions of acetylenic hydrogen withbromoalkenes, iodoarenes, and bromopyridines. Tetrahedron Lett1975;50:4467–70.

[25] Gasparini F, Andres H, Flor PJ, et al. [(3)H]-M-MPEP, a potent,subtype-selective radioligand for the metabotropic glutamate receptorsubtype 5. Bioorg Med Chem Lett 2002;12:407–9.

[26] Dollé F. [18F]Fluoropyridines: from conventional radiotracers to thelabelling of macromolecules such as proteins and oligonucleotides.PET Chemistry: the driving force in molecular imaging, Ernst ScheringResearch Foundation. Berlin: Springer; 2007. p. 113–57.

[27] Wilson AA, Houle S. Radiosynthesis of carbon-11 labeled N-methyl-2-(arylthio)benzylamines: potential racers for the serotonin reuptakereceptor. J Labelled Compd Radiopharm 1999;42:1277–88.