Development of N-[3-(29,49-Dichlorophenoxy)-2-18F-Fluoropropyl]-N-Methylpropargylamine (18F-Fluoroclorgyline)

as a Potential PET Radiotracer for Monoamine Oxidase-AJogeshwar Mukherjee and Zhi-Ying Yang

FRANKLIN MCLEAN INSTITUTE, DEPARTMENT OF RADIOLOGY AND DEPARTMENT OF PHARMACOLOGICAL AND

PHYSIOLOGICAL SCIENCES, UNIVERSITY OF CHICAGO, CHICAGO, ILLINOIS, USA

ABSTRACT. We have synthesized N-[3-(2*,4*-dichlorophenoxy)-2-18F-fluoropropyl]-N-methylpropar-gylamine (18F-fluoroclorgyline) as a potential positron emission tomography (PET) radiotracer for mono-amine oxidase A (MAO-A). The radiosynthesis was carried out by a 18F-fluoride-for-mesylate substitutionreaction in approximately 20% radiochemical yield in specific activities of 1–2 Ci/mmol. Selectivity forMAO-A was demonstrated by the high affinity of clorgyline (IC50 5 39 nM) and lower affinity of(R)-deprenyl (IC50 > 100 mM) for the inhibition of 18F-fluoroclorgyline binding in vitro in rat brainmembranes. The uptake of 18F-fluoroclorgyline in the rat brains was high (>1.0% injected dose/g). Thebinding of 18F-fluoroclorgyline in the rat brain correlated with the distribution of MAO-A and was inhibitedby preadministration of MAO-A inhibitors, clorgyline, and Ro 41-1049, whereas (R)-deprenyl, a MAO-Bblocker, had no inhibitory effect. These results suggest that 18F-fluoroclorgyline is a potential PETradiotracer for MAO-A. NUCL MED BIOL 26;6:619–625, 1999. © 1999 Elsevier Science Inc. All rightsreserved.

KEY WORDS. 18F-Fluoroclorgyline, Monoamine oxidase A , Fluorine-18

INTRODUCTION

Monoamine oxidase A (MAO-A) has been investigated noninva-sively by using radioligands developed for positron emission tomog-raphy (PET) and single photon emission computed tomography(SPECT). Carbon-11-labeled irreversible radiotracers, such as 11C-clorgyline has been extensively studied in humans (3, 8). Effortshave also been made to develop iodinated and fluorinated analogsof clorgyline, which may provide potential radioligands for the studyof this enzyme system (Fig. 1). Other radiotracers for MAO-A, suchas 11C-harmine and derivatives and 11C-brofaromine (1, 2), havealso been investigated more recently for use in PET.

Our interest was to develop a fluorine-18 radiotracer for the studyof MAO-A to use in the evaluation of various antidepressants fortheir affinity to the enzyme in vitro and their ability to inhibit theenzyme in vivo. The longer half-life (110 min) of fluorine-18 wouldallow the in vitro and in vivo pharmacological evaluation of thevarious drugs over an extended period of time compared withcarbon-11 (half-life of 20 min). A brief structure–activity relation-ship of clorgyline has been carried out with iodine, fluorine andphenoxy substitutions (11, 13, 14). A fluorine-18-labeled analog ofclorgyline, using the lower specific activity electrophilic fluorine-18, has been reported previously and shown to localize in MAO-Asites in rodents (5).

We decided on the development of irreversible agents due to thesomewhat lesser synthetic and radiosynthetic difficulties posed bythese compounds. There are only limited structure–activity rela-tionship data available on fluorinated derivatives of clorgyline (13).

A 11C-radiolabeled tracer of an aromatic ring fluorinated derivativeshowed promise in localizing to MAO-A sites in mice (14). Apreliminary report of a low specific activity electrophilic fluorine-18-labeled clorgyline derivative with promising properties has beenreported (5). Due to the general difficulty of radiolabeling withnucleophilic fluorine-18 on electron-rich aromatic rings in order toobtain a high specific activity fluorine-18 radiotracer, we investi-gated other sites for incorporating the fluorine atom. We envisionedplacing the fluorine on the propyl group, at the b-carbon ofclorgyline, which would potentially provide a fluorinated MAO-Ainhibitor that could be developed into a 18F-fluorinated derivativeas a potential PET radiotracer. We report here the radiosynthesisand in vitro and in vivo characterization of N-[3-(29,49-dichlorophenoxy)-2-18F-fluoropropyl]-N-methylpropargylamine(18F-fluoroclorgyline) as a potential PET radiotracer for MAO-A.

MATERIALS AND METHODS

N-Methylpropargylamine, 2,4-dichlorophenol, and 1,3-dibromo-2-propanol were purchased from Aldrich Chemical Co. (Milwaukee,WI) and clorgyline, Ro 41-1049 (N-(2-aminoethyl)-5-(3-fluoro-phenyl)-4-thiazolecarboxamide) and (R)-deprenyl were purchasedfrom Research Biochemicals Inc. (Natick, MA). All other reagentswere of analytical grade and were used without further purification.Analytical and preparative thin-layer chromatography (TLC) wereperformed using Baker-flex silica gel IB-F and Alltech DC-Fertig-platten SIL G-200 UV254 plates, respectively. High pressure liquidchromatography (HPLC) was carried out on a Gilson GradientSystem consisting of two Gilson pumps and one ultraviolet (UV)detector with wavelength fixed at 254 nm and a radiation flowdetector with a NaI(Tl) crystal. Solvent A of the gradient was0.01 M phosphoric acid, pH 4.7, and solvent B was acetonitrile.Semi-prep (250 3 10 mm) C18 columns from Alltech Associates

Address correspondence to: Jogeshwar Mukherjee, Ph.D., Department ofNuclear Medicine/PET, Kettering Medical Center, 3535 Southern Boulevard,Dayton, OH 45429, USA; e-mail: jogeshwar_mukherjee@ketthealth.com

Received 8 August 1998.Accepted 5 April 1999.

Nuclear Medicine & Biology, Vol. 26, pp. 619–625, 1999 ISSN 0969-8051/99/$–see front matterCopyright © 1999 Elsevier Science Inc. All rights reserved. PII S0969-8051(99)00027-X

Inc. (Deerfield, IL) were used for reverse-phase HPLC. Protonnuclear magnetic resonance (NMR) spectra were obtained on a GENMR OMEGA 500 MHz spectrometer, and electron-impact massspectra were obtained on a VG Instruments Inc. Model 7250 massspectrometer.

Synthesis

3-(2*,4*-DICHLOROPHENOXY)-2-HYDROXYPROPYL BROMIDE (1). Around-bottomed flask was charged with 2,4-dichlorophenol (0.16 g;

1 mmol) and dissolved in 1 mL of dimethylacetamide (DMA). Intothis solution was added sodium hydride (0.05 g; 2 mmol), and thesolution was heated at 50°C for 30 min. Into this reaction mixture,1,3-dibromo-2-propanol (0.44 g; 2 mmol) was added and thereaction was heated overnight (20 h) at 70°C (reaction schemeshown in Fig. 2) and the product appeared with an Rf 5 0.6(3:1 hexane-ethylacetate). The reaction mixture was subsequentlyquenched with water and the product was extracted with ether andpurified on preparative TLC (3:1 hexane-ethylacetate) to provide3-(29,49-dichlorophenoxy)-2-hydroxypropyl bromide in 45% yield.

FIG. 1. Chemical structures ofradiolabeled clorgyline analogsas monoamine oxidase A inhibi-tors.

620 J. Mukherjee and Z.-Y. Yang

NMR d ppm: 7.37 (d, 1H, Ar-3H), 7.16 (dd, 1H, Ar-5H), 6.94 (d,1H, Ar-6H), 4.05 (m, 1H, CH), 3.58 (d, 2H, OCH2), 3.42 (d, 2H,CH2Br).

N - [ 3 - ( 2 * , 4 * - D I C H L O R O P H E N O X Y ) - 2 - H Y D R O X Y P R O P Y L ] - N -

METHYLPROPARGYLAMINE (2). Alkylation was carried out by usingthe reported general method of alkylation of propargylamines (12,17). In a round-bottomed flask, N-methylpropargylamine (69 mg,1 mmol) was taken along with 3-(29,49-dichlorophenoxy)-2-hy-droxypropyl bromide (0.30 g, 1 mmol) in triethylamine (100 mL)and this mixture was refluxed for 2 h. The product appeared with anRf of 0.35 (5% CH3OH, 95% CH2Cl2). The mixture was purifiedon preparative TLC to provide the pure product, N-[3-(29,49-dichlorophenoxy)-2-hydroxypropyl]-N-methylpropargylamine in60% yield and used subsequently for mesylation. NMR d ppm: 7.28(d, 1H, Ar-3H), 7.10 (dd, 1H, Ar-5H), 6.8 (d, 1H, Ar-6H), 4.08 (d,2H, OCH2), 3.72 (s, 2H, -NCH2CCH), 2.78 (s, 3H, NCH3).

N - [ 3 - ( 2 * , 4 * - D I C H L O R O P H E N O X Y ) - 2 - M E S Y L O X Y P R O P Y L ] - N -

METHYLPROPARGYLAMINE (3). The alcohol, N-[3-(29,49-dichloro-phenoxy)-2-hydroxypropyl]-N-methylpropargylamine (29 mg; 0.1mmol) was taken in dichloromethane (2 mL) and pyridine (10 mL)and the mixture was cooled in an ice-water bath. Into this cooled

solution was added methanesulfonyl chloride (10 mL, 0.1 mmol)and the reaction was stirred for 2 h. The product appeared with anRf of 0.80 (5% CH3OH, 95% CH2Cl2). The reaction mixture wasdried in vacuo, taken up in dichloromethane, and washed withwater. The dried organic layer was subsequently purified on prepar-ative TLC to provide the pure product, N-[3-(29,49-dichlorophe-noxy)-2-mesyloxypropyl]-N-methylpropargylamine in 55% yield.NMR d ppm: 7.35 (d, 1H, Ar-3H), 7.18 (dd, 1H, Ar-5H), 7.0 (d,1H, Ar-6H), 4.04 (d, 2H, OCH2), 3.76 (s, 2H, -NCH2CCH), 3.62(s, 3H, CH3SO2), 2.73 (s, 3H, NCH3); Mass spectra (m/z, %) 366(M1, 4%).

N-[3-(2*,4*-DICHLOROPHENOXY)-2-FLUOROPROPYL]-N-METHYLPROPAR-

GYLAMINE (4). The mesylate, N-[3-(29,49-dichlorophenoxy)-2-me-syloxypropyl]-N-methylpropargylamine, 3 (36 mg; 0.1 mmol) wastaken in tetrahydrofuran (0.2 mL). Into this mixture was addedtetrabutylammonium fluoride (TBAF; 0.1 mL of a 1-M tetrahydro-furan solution) and the reaction was refluxed for 45 min. Thereaction mixture was subsequently taken to dryness, the residuetaken up in CH2Cl2 and washed with water and dried overanhydrous magnesium sulfate. The product appeared with an Rf of0.45 (5% CH3OH, 95% CH2Cl2). The dried organic layer was

FIG. 2. Synthesis scheme of the mesylate precursor 3 (N-[3-(2*,4*-dichlorophenoxy)-2-mesyloxypropyl]-N-methylpropar-gylamine, N-[3-(2*,4*-dichlorophenoxy)-2-fluoropropyl]-N-methylpropargylamine 4 (fluoroclorgyline), and N-[3-(2*,4*-dichloro-phenoxy)-2-18F-fluoropropyl]-N-methylpro- pargylamine 18F-4 (18F-fluoroclorgyline).

18F-Fluoroclorgyline as a Radiotracer for Monoamine Oxidase A 621

subsequently purified on preparative TLC to provide the pureproduct, N-[3-(29,49-dichlorophenoxy)-2-fluoropropyl]-N-methyl-propargylamine in 30% yield. NMR d ppm: 7.38 (d, 1H, Ar-3H),7.12 (dd, 1H, Ar-5H), 6.90 (d, 1H, Ar-6H), 4.10 (d, 2H, OCH2),3.46 (s, 2H, -NCH2CCH), 2.80 (s, 3H, NCH3); Mass spectra (m/z,%) 290 (M1, 8%).

N - [ 3 - ( 2 * , 4 * - D I C H L O R O P H E N O X Y ) - 2 - 1 8 F - F L U O R O P R O P Y L ] - N -

METHYLPROPARGYLAMINE (18F-FLUOROCLORGYLINE, 18F-4). Aqueous18F-fluoride (10–100 mCi) was treated with potassium carbonate(0.34 mg) and Kryptofix (1.88 mg) in a Pyrex round-bottomed flask.The mixture was dried by heating at 85–90°C under a stream ofnitrogen. Last traces of water were removed azeotropically withacetonitrile (1–2 mL). Into this dried Kryptofix-K18F complex, asolution (2 mg in 200 mL of acetonitrile) of N-[3-(29,49-dichloro-phenoxy)-2-mesyloxypropyl]-N-methylpropargylamine was added,and this mixture was refluxed at 80–85°C. After 20 min, theacetonitrile in the mixture was dried under nitrogen and the residuewas taken up in methanol. This methanolic solution was passedthrough a mini-column of neutral alumina, followed by washing thealumina with methanol. The methanolic eluate was evaporated tonear dryness and the residue taken up in 60% methanol/40% buffer(,1 mL) and injected into the HPLC. Separation was carried outon an Alltech C18 column, using gradient elution with 0.01 Mphosphoric acid and acetonitrile (95% 0.01 M phosphoric acid and5% acetonitrile at 0 min, flow rate of 3 mL/min to 5% 0.01 Mphosphoric acid and 60% acetonitrile at 20 min, flow rate of3 mL/min). The retention time of the radiotracer, N-[3-(29,49-dichlorophenoxy)-2-18F-fluoropropyl]-N-methylpropargylamine(18F-fluoroclorgyline) was 16.2 min, whereas that of the mesylateprecursor was approximately 17.5 min (Fig. 3). The yields variedbetween 15% and 20% (decay-corrected) in several experimentswith a radiochemical purity .95%.

To determine specific activity of 18F-fluoroclorgyline, the frac-tion collected between 15.6 and 16.5 min (containing the majorityof the radioactive peak eluting at 16.2 min, as seen in Fig. 3) in thefirst HPLC purification was reinjected under the same gradientconditions for a second purification. A fraction between 15.6 and16.5 min was collected to provide 18F-fluoroclorgyline and all themass associated with this peak was assumed to be that of 18F-fluoroclorgyline for purposes of computing specific activity. Thismass peak was compared with standard fluoroclorgyline injections.

The specific activity of HPLC purified 18F-fluoroclorgyline variedbetween 1 and 2 Ci/mmol.

In Vitro Binding Assays

Rat brains were isolated and homogenized with a Tekmar Tissum-izer (15 s at half-maximum speed) in a 100-fold (w:v) dilution of a50 mM Tris-HCl buffer, pH 7.4, containing 120 mM NaCl, 5 mMKCl, 2 mM CaCl2, 1 mM MgCl2, 1 mM Na-ethylene diaminetetraacetic acid (Na-EDTA), and 0.1 mM Na ascorbate. Thehomogenate was centrifuged at 12,000 3 g for 15 min at 4°C. Thepellet was resuspended in the same volume of buffer, centrifuged asecond time, and resuspended in fresh buffer at a concentration of50 mg of tissue/mL. Each assay tube contained 0.10 mL of this stocksolution.

In vitro binding affinities of clorgyline, Ro 41-1049 and (R)-deprenyl to MAO-A in rat (Sprague–Dawley) brain homogenateswas carried out by incubating various concentrations (0.01 nM to0.1 mM) of compounds along with the radioligand, 18F-fluoroclor-gyline (specific activity 1 Ci/mmol). Binding was initiated byaddition of the tissue homogenate, and the tubes were incubated for1 h at 37°C. The binding was terminated by filtration using aBrandel filtration apparatus, followed by washings with cold 50 mMTris-HCl buffer (4 3 1 mL). Nonspecific binding was determined inthe presence of 10 mM clorgyline. The filters were counted in awell-counter for fluorine-18 activity. The data were analyzed usingthe Ligand program and IC50 values for the various drugs wereobtained (9). The binding curves were displayed using GraFit.

In Vivo Studies

For in vivo studies, groups (n 5 4) of Sprague–Dawley rats (250 g)were administered clorgyline (10 mg/kg), Ro 41-1049 (10 mg/kg),or (R)-deprenyl (10 mg/kg). All compounds (including saline, forcontrol rats) were administered intraperitoneally under anesthesia,90 min before injection of the radioligand and the rats were allowedfree access to food and water during the interval. The radioligand,18F-fluoroclorgyline, 100 mCi (specific activity 1 Ci/mmol), wasadministered intravenously into each rat under anesthesia. The ratswere subsequently allowed to recover and had free access to foodand water. All rats were sacrificed 75 min after the radioligandinjection and the various brain regions (cerebrum and cerebellum)

FIG. 3. High performance liquid chro-matography (HPLC) chromatogram ofN-[3-(2*,4*-dichlorophenoxy)-2-18F-fluoropropyl]-N- methylpropar-gylamine (18F-fluoroclorgyline). Thex-axis is time in minutes and y-axis isabsorbance at 254 nm. Broken line isthe radioactivity detector and the solidline is absorbance.

622 J. Mukherjee and Z.-Y. Yang

and blood were isolated into tared vials and counted for fluorine-18activity to provide a percent of injected dose of 18F-fluoroclorgy-line/g of wet tissue for each group of rats. Metabolite analysis wascarried out in the blood plasma (75 min postinjection of 100 mCi of18F-fluoroclorgyline). The blood samples (1 mL) were centrifuged at1,000 3 g for 4 min to separate plasma from blood cells and theplasma was removed for further analyses. The plasma (200 mL) wascounted for total radioactivity before addition of 100 mL of sodiumbicarbonate. The plasma mixture was then extracted with ethylacetate (3 3 500 mL) and the combined ethyl acetate extracts werefiltered and evaporated to dryness, and the residue was redissolvedin methanol for TLC analysis. The aqueous plasma layer andcombined ethyl acetate extracts were also counted in a well-counterto evaluate percent of parent remaining in the plasma. The TLCplates were developed in chloroform-methanol (9:1) and TLCanalysis was performed and compared with an authentic sample of18F-fluoroclorgyline (Rf 5 0.70 for 18F-fluoroclorgyline), and cutinto 1-cm wide strips, which were then counted in a well counter.Similarly, the excised brain regions were homogenized with 5%sodium bicarbonate (500 mL) and centrifuged at 1,000 3 g for10 min. The aqueous layer was removed and extracted three timeswith ethyl acetate (500 mL each time). The solutions were countedand analyzed by TLC and the Rfs compared with an authenticsample of 18F-fluoroclorgyline as described above, indicating theradioactivity to be primarily the parent radioligand.

RESULTS

The synthesis of the precursor, N-[3-(29,49-dichlorophenoxy)-2-mesyloxypropyl]-N-methylpropargylamine, was carried out inmoderate yields in a three-step reaction starting from 2,4-dichloro-phenol and 1,3-dibromo-2-propanol (reaction scheme shown in Fig.2). The coupling of the sodium dichlorophenoxide with 1,3-dibromo-2-propanol provided the desired product in approximately40–45% yields. Subsequently, this product was coupled to N-methylpropargylamine to provide the desired alcohol, N-[3-(29,49-dichlorophenoxy)-2-hydroxypropyl]-N-methylpropargylamine inapproximately 60% yields. Treatment of the mesylate with TBAFusing reported procedures provided the unlabeled fluorinated deriv-ative, N-[3-(29,49-dichlorophenoxy)-2-fluoropropyl]-N-methylpro-pargylamine (fluoroclorgyline). For radiolabeling purposes the alco-hol had to be converted to the corresponding tosylate or mesylate.Tosylation of the alcohol did not provide significant amounts of theproduct. It was concluded that the steric effects might be playing arole in inhibiting the tosylation reaction. Subsequently, with theless bulkier group, mesylation led to the formation of the desiredproduct. The product was found to be quite stable for use inradiolabeling reactions. The radiosynthesis of 18F-fluoroclorgylineusing a 18F-fluoride-for-mesylate nucleophilic exchange proceededsmoothly with high radiochemical yields. The mixture was purifiedon reverse-phase HPLC to provide high specific activities of theradiolabeled product. The desired product, 18F-fluoroclorgyline,eluted at approximately 16.2 min (Fig. 3).

A significantly large radiolabeled side-product, in radiochemicalyields ranging between 50% and 80% of the product peak, wasobserved that eluted before 18F-fluoroclorgyline, at approximately14.5 min as shown in the HPLC chromatogram in Figure 3. Thisunidentified radiolabeled side-product did not bind to MAO-A. Itis hypothesized that under the reaction conditions, the mesylateprecursor may lend itself to a nucleophilic substitution reaction bythe b-nitrogen to form an “aziridinium intermediate.” Nucleophilic18F-fluoride may thus potentially interact with the aziridinium

intermediate in two ways: (a) substitution at the tertiary carbon toprovide the desired product, 18F-fluoroclorgyline and (b) substitu-tion at the secondary carbon to provide the side-product.

The in vitro binding profiles of clorgyline, Ro 41-1049, and(R)-deprenyl are shown in Figure 4. Clorgyline, which is a MAO-Aselective agent (6), exhibited a high affinity (IC50 5 39 nM) andRo 41-1049 exhibited lower affinity than clorgyline (IC50 5420 nM). As expected, (R)-deprenyl, which is a potent MAO-Binhibitor (7), showed very weak binding (IC50 $ 100 mM). Thisbinding is indicative of the selective labeling of MAO-A sites by

FIG. 4. In vitro binding of the compounds measured by using18F-fluoroclorgyline in rat brain homogenates.

FIG. 5. In vivo binding of 18F-fluoroclorgyline in the ratbrains (cerebrum and cerebellum) and blood are shown aspercent of injected dose/g of wet tissue. Rats were pretreatedeither with saline (control) or the various drugs, clorgyline(10 mg/kg), Ro-41-1049 (10 mg/kg), and (R)-deprenyl(10 mg/kg) administered intraperitoneally 90 min before theradioligand. All rats were sacrificed 75 min after intravenousinjection of the radioligand, 18F-fluoroclorgyline.

18F-Fluoroclorgyline as a Radiotracer for Monoamine Oxidase A 623

the radioligand, 18F-fluoroclorgyline, which is similar to the re-ported analogs of clorgyline (13).

The in vivo effects of the compounds on the binding of 18F-fluoroclorgyline in the rat brains are shown in Figure 5. Binding wasseen both in the cerebrum and cerebellum, with the former showinggreater binding. The binding was reduced significantly by bothMAO-A inhibitors, clorgyline, and Ro 41-1049. Clorgyline reducedthe binding to less than 40% of the controls, whereas R0 41-1049reduced the binding to less than 50% of the controls. The effect ofthe MAO-B inhibitor, (R)-deprenyl, on the binding of 18F-fluoro-clorgyline was absent, consistent with the poor affinity of (R)-deprenyl for MAO-A sites labeled with 18F-fluoroclorgyline in vitro.Metabolite analysis in the plasma indicated the presence of anunidentified hydrophylic component as well as the parent, 18F-fluoroclorgyline. At 75 min postinjection of 18F-fluoroclorgyline,approximately 25% of the radioactivity was found to be unchangedradiotracer in the plasma, whereas 75% of 18F-fluoroclorgyline wasthe hydrophilic metabolite. Ethyl acetate extracts of homogenizedbrain, 75 min postinjection of 18F-fluoroclorgyline, indicated .90%18F-fluoroclorgyline with trace amounts of lipophilic constituents.

DISCUSSION

We have developed a fluorine-18 analog of the irreversible inhib-itor, clorgyline, to study the biodistribution of MAO-A. Using thelimited structure–activity data on clorgyline analogs, we included afluorine at the b-carbon in the propyl group of clorgyline, such thata fluorine-18 analog could be readily prepared by nucleophilicsubstitution of an appropriate leaving group. Although placing ofthe fluorine at the b-position was expected to cause a lowering ofthe basicity of the propargyl nitrogen, mechanistic implications ofthis on the affinity were not clear, because the nitrogen may notdirectly interact with the binding site (16). The synthesis proceededsmoothly to provide the precursor mesylate in three steps, startingfrom 2,4-dichlorophenol. Initial attempts to synthesize the corre-sponding tosylate precursor did not procced well, probably due tosteric constraints on the hydroxyl group. The radiolabeling of themesylate to provide 18F-fluoroclorgyline proceeded in moderateyields, which was purified on reverse-phase HPLC for use in in vitroand in vivo studies. It should be noted that all studies were carriedout with the racemic mixture, because the b-position is a chiralcenter.

Affinities of the MAO-A inhibitors, clorgyline, and Ro 41-1049for the radioligand, 18F-fluoroclorgyline, were found to be 39 and420 nM, respectively. These findings are in agreement with thereported relative affinities of these two compounds (15). Theaffinity of (R)-deprenyl, the MAO-B inhibitor, was significantlyweak (IC50 $ 100 mM), indicative of the good selectivity of18F-fluoroclorgyline for MAO-A sites.

Distribution studies of 18F-fluoroclorgyline revealed interestingfindings, somewhat similar to previous reports (10). Binding of18F-fluoroclorgyline in various brain regions after intravenous ad-ministration to rats was observed. Cerebrum and cerebellum showedapproximately similar amounts of binding. Preadministration ofclorgyline and Ro 41-1049 to the rats reduced the amount ofbinding of 18F-fluoroclorgyline in the various brain regions. Ratiosof control versus clorgyline treated rats were approximately 2.5 incerebrum and cerebellum, which is similar to that reported for11C-clorgyline in mice (8). Inhibitory effects of Ro 41-1049 wereslightly lower than those of clorgyline, which is in agreement withthe observed lower binding affinity of Ro 41-1049 than clorgyline.No effect was observed in rats pretreated with (R)-deprenyl.

Although the ratio of specific-to-nonspecific binding of 18F-fluoroclorgyline was found to be similar to that of 11C-clorgyline,the longer half-life of fluorine-18 provides provides the ability toprolong the in vivo experiments to study the slow binding kinetics ofthe irreversible inhibitors. Furthermore, the longer half-life offluorine-18 makes in vitro binding experiments possible, thus en-abling their correlation with in vivo findings.

Radioactivity present in the blood as seen in Figure 5 is due to thepresence of 18F-fluoroclorgyline as well as a hydrophilic metabolite.It has been suggested that the alkylated propargylamines may leadto aliphatic acids as metabolites if they follow a similar metabolicpattern as deprenyl, that is, cleavage of the propargyl moiety,followed by oxidative deamination of the alkylamines (18). Basedon our findings, it is conceivable that this mechanism may beoperative in the case of 18F-fluoroclorgyline to result in thehydrophilic metabolite. Because MAO inhibition is one of themechanisms of antidepressant therapy, 18F-fluoroclorgyline maypotentially be a fluorinated PET radiotracer for future studies inanimal models for drug evaluations.

CONCLUSION

The radiosynthesis of 18F-fluoroclorgyline was achieved in moderateradiochemical yield using a one-step direct nucleophilic displace-ment reaction. Binding of 18F-fluoroclorgyline showed a more than1,000-fold selectivity for MAO-A over that of MAO-B, observedusing deprenyl and clorgyline. In the rodent brain, 18F-fluoroclor-gyline accumulated in areas known to contain significant amountsof MAO-A. The binding of 18F-fluoroclorgyline to MAO-A in vivowas blocked by pretreatment with clorgyline and Ro 41-1049. Theseresults suggests that 18F-fluoroclorgyline is a potentially useful PETradiotracer for MAO-A.

Financial support was provided by U.S. Department of Energy DE-FG02-98ER62540. We like to thank Robert Mintzer and Dr. BingzhiShi for technical assistance.

References1. Bergstrom M., Westerberg G., Kihlberg T. and Langstrom B. (1997)

Synthesis of some 11C-labeled MAO-A inhibitors and their in vivouptake kinetics in rhesus monkey brain. Nucl. Med. Biol. 24, 381–388.

2. Bergstrom M., Westerberg G. and Langstrom B. (1997) 11C-Harmine asa tracer for monoamine oxidase A (MAO-A): In vitro and in vivostudies. Nucl. Med. Biol. 24, 287–293.

3. Fowler J. S., MacGregor R. R., Wolf A. P., Arnett C. D., Dewey S. L.,Schlyer D., Christman D., Logan J., Smith M. and Sachs H. (1987)Mapping human brain monoamine oxidase A and B with 11C-labeledsuicide inactivators and PET. Science 235, 481–485.

4. Hirata M., Magata Y., Ohmono Y., Saji H., Murakami K., Takagaki T.,Yamamura N., Tanaka C., Konishi J. and Yokoyama A. (1995)Evaluation of radioiodinated iodoclorgyline as a SPECT radiopharma-ceutical for MAO-A in the brain. Nucl. Med. Biol. 22, 175–180.

5. Hirata M., Magata Y., Ohmono Y., Saji H., Tanaka C. and YokoyamaA. (1994) Radiofluorinated clorgyline derivative for mapping MAO-Aactivity in brain with PET. J. Labelled Compd. Radiopharm. 35, 234–235.

6. Johnston J. P. (1968) Some observations upon a new inhibitor ofmonoamine oxidase in brain tissue. Biochem. Pharmacol. 17, 1285–1297.

7. Knoll J. and Magyar K. (1972) Some puzzling pharmacological effects ofmonoamine oxidase inhibitors. Adv. Biochem. Psychopharmacol. 5,393–408.

8. MacGregor R. R., Halldin C., Fowler J. S., Wolf A. P., Arnett C. D.,Langstrom B. and Alexoff D. (1985) Selective, irreversible in vivobinding of [11C]clorgyline and [11C]-L-deprenyl in mice: Potentialfor measurement of functional monoamine oxidase activity in brainusing positron emission tomography. Biochem. Pharmacol. 34, 3207–3210.

9. Munson P. J. and Rodbard D. (1980) LIGAND: A versatile approach

624 J. Mukherjee and Z.-Y. Yang

for the characterisation of ligand binding systems. Anal. Biochem. 107,220–239.

10. Murakami M., Kondoh Y., Weimin Y., Mizusawa S., Nakamichi H.,Takahashi K., Sasaki H., Iida H., Miura S., Kanno I. and Uemura K. (1992)A convenient method for regional monoamine oxidase-A determinationby 14C-clorgyline autoradiography. Nucl. Med. Biol. 19, 619–626.

11. O’Brien E. M., Tipton K. F., Meroni M. and Dostert P. (1994)Inhibition of monoamine oxidase by clorgyline analogues. J. NeuralTrans. 41(suppl.), 295–305.

12. Ohmono Y., Hirata M., Murakami K., Magata Y., Tanaka C. andYokoyama A. (1991) Synthesis of fluorine and iodine analogues ofclorgyline and selective inhibition of monoamine oxidase A. Chem.Pharm. Bull. 39, 1038–1040.

13. Ohmono Y., Hirata M., Murakami K., Magata Y., Tanaka C. andYokoyama A. (1991) Synthesis of [125I]iodoclorgyline, a selectivemonoamine oxidase A inhibitor, and its biodistribution in mice. Chem.Pharm. Bull. 39, 3343–3345.

14. Ohmono Y., Hirata M., Murakami K., Magata Y., Tanaka C. andYokoyama A. (1993) Synthesis and characterization of [11C]fluoroclo-rgyline: A monoamine oxidase A specific inhibitor for positron emissiontomography. Chem. Pharm. Bull. 41, 1994–1997.

15. Saura J., Kettler R., Da Prada M. and Richards J. G. (1992)Quantitative enzyme radioautography with [3H]Ro 41-1049 and[3H]Ro 19-6327 in vitro: Localization and abundance of MAO-A andMAO-B in rat CNS, peripheral organs, and human brain. J. Neuro-sci. 12, 1977–1999.

16. Walsh C. (1979) Enzymatic Reaction Mechanisms, pp. 402–405. W.H.Freeman and Company, San Francisco.

17. Yu P. H., Davies B. A. and Boulton A. A. (1992) Aliphatic propar-gylamines: Potent, selective, irreversible monoamine oxidase B inhibi-tors. J. Med. Chem. 35, 3705–3713.

18. Yu P. H., Davies B. A. and Boulton A. A. (1993) Effect of structuralmodification of alkyl N- propargylamines on the selective inhibition ofmonoamine oxidase B activity. Biochem. Pharmacol. 46, 753–757.

18F-Fluoroclorgyline as a Radiotracer for Monoamine Oxidase A 625