Available online at www.sciencedirect.com

Nuclear Medicine and Biology 38 (2011) 137–148www.elsevier.com/locate/nucmedbio

A rapid solid-phase extraction method for measurement ofnon-metabolised peripheral benzodiazepine receptor ligands,[18F]PBR102 and [18F]PBR111, in rat and primate plasma

Andrew Katsifisa,⁎, Christian Loc'ha, David Hendersonb, Thomas Bourdiera, Tien Phama,Ivan Gregurica, Peter Lamb, Paul Callaghana, Filomena Mattnera,

Stefan Eberlb,c, Michael Fulhamb,c,d

aANSTO LifeSciences, Sydney, 2234, AustraliabDepartment of PET and Nuclear Medicine, Royal Prince Alfred Hospital, Sydney, 2050 Australia

cSchool of Information Technology, University of Sydney, Sydney, 2006 AustraliadSydney Medical School, University of Sydney, Sydney, 2006 Australia

Received 10 May 2010; received in revised form 21 July 2010; accepted 25 July 2010

Abstract

Objectives: To develop a rapid and reliable method for estimating non-metabolised PBR ligands fluoroethoxy ([18F]PBR102)- andfluoropropoxy ([18F]PBR111)-substituted 2-(6-chloro-2-phenyl)imidazo[1,2-a]pyridine-3-yl)-N,N-diethylacetamides in plasma.Methods: Rats and baboons were imaged with PET up to 2 h postinjection of [18F]PBR102 and [18F]PBR111 under baseline conditions,after pre-blocking or displacement with PK11195. Arterial plasma samples were directly analysed by reverse-phase solid-phase extraction(RP-SPE) and RP-HPLC and by normal-phase TLC. SPE cartridges were successively washed with acetonitrile/water mixtures. SPE eluantradioactivity was measured in a γ-counter to determine the parent compound fraction and then analysed by HPLC and TLC for validation.Results: In SPE, hydrophilic and lipophilic radiolabelled metabolites were eluted in water and 20% acetonitrile/water. All non-metabolised[18F]PBR102 and [18F]PBR111 were in SPE acetonitrile fraction as confirmed by HPLC and TLC analysis. Unchanged (%) [18F]PBR102and [18F]PBR111 from SPE analysis in rat and baboon plasma agreed with those from HPLC and TLC analysis. In rats and baboons, thefraction of unchanged tracer followed a bi-exponential decrease, with half-lives of 7 to 10 min for the fast component and N80 min for theslow component for both tracers.Conclusions: Direct plasma SPE analysis of [18F]PBR102 and [18F]PBR111 can reliably estimate parent compound fraction. SPE wassuperior to HPLC for samples with low activity; it allows rapid and accurate metabolite analysis of a large number of plasma samples forimproved estimation of metabolite-corrected input function during quantitative PET imaging studies.Crown Copyright © 2011 Published by Elsevier Inc. All rights reserved.

Keywords: [18F]PBR102; [18F]PBR111; Metabolite analysis; SPE; HPLC; TLC

1. Introduction

The peripheral benzodiazepine receptor (PBR), recentlyrenamed translocator protein TSPO 18 kDa, is a transmem-brane multimeric protein complex primarily located in theouter mitochondrial membrane. The PBR is predominantly

⁎ Corresponding author. ANSTO LifeSciences, Australian NuclearScience and Technology Organisation, Kirrawee NSW, 2232, Australia.Tel.: +61 2 9717 9094; fax: +61 2 9717 9262.

E-mail address: andrew.katsifis@ansto.gov.au (A. Katsifis).

0969-8051/$ – see front matter. Crown Copyright © 2011 Published by Elsevierdoi:10.1016/j.nucmedbio.2010.07.008

expressed in peripheral organs such as the kidney, heart [1]and the steroid hormone-producing cells of the adrenalcortex, testis and ovary [2]. Changes in PBR density havebeen observed in a number of pathological conditions, inparticular, in activated microglia after neuronal damage orneurodegeneration and in inflammation [3,4]. These findingssuggest that radioligands for the PBRs could be usefulmarkers for the detection of such disorders using PET andSPECT [5]. In prior work, [11C]PK11195 has been used tomap PBR in the human heart, brain and endocrine tissue[6,7], and also for cerebral inflammation [8–11].

Inc. All rights reserved.

Fig. 1. [18F]PBR102 and [18F]PBR111 chemical structures.

138 A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

The increasing availability of PET and the limitations ofcarbon-11 labelled PBR compounds have prompted thedevelopment of novel [18F]-labelled PBR ligands [12,13].We recently reported our development of the fluoroethoxy-and fluoropropoxy-substituted 2-(6-chloro-2-phenyl)-imidazo[1,2-a]pyridine-3-yl)-N,N-diethylacetamides [18F]PBR102and [18F]PBR111 to image PBR expression [14,15].

Quantification of cerebral receptor binding sites in vivorequires the application of appropriate multi-compartmentalkinetic models that describe the ligand–receptor interaction.Such models also require an input function which representsthe amount of unchanged tracer in plasma over time. Toestimate the amount of non-metabolised tracer and providemetabolite-corrected input functions for quantitative PETimaging in rats and primates, a rapid and reliable method formetabolite analysis of [18F]PBR102 and [18F]PBR111 inplasma is required.

Several chromatography techniques to measure unchangedradiopharmaceuticals have been reported and their advantagesdiscussed by Maziere et al [16]. Normal- and reverse-phasethin layer chromatography (radio-TLC) is a long procedure,and the quantification of multiple and/or minor radio-labelledmetabolites remains difficult. High-performance liquid chro-matography (radio-HPLC) is considered to be the referencemethod as it allows an accurate determination of unchangedtracers in the presence of hydrophilic and lipophilic radio-labelled metabolites when using appropriate reverse-phasecolumns and solvents. The necessary elimination of plasmaprotein before standard HPLC, however, is the majordrawback of this technique. Protein precipitation using acidsor organic solvents prior to HPLC analysis is time consuming,and the recovery of the entire radioactivity in the supernatant isdifficult when metabolites and parent compounds are verylipophilic. These drawbacks have been overcome by usinginternal-surface reversed-phase HPLC columns for extractionof lipophilic metabolites and the parent compound fromplasma [17]. Unfortunately, these columns are unstable andexpensive for routine use. The switching valve procedure[18–20] and a dedicated HPLC pre-column for proteinelimination have improved the routine use of HPLC inmetabolite determination [21,22].

Liquid–liquid extraction and solid-phase extraction(SPE) facilitate radiopharmaceutical preparation by elimi-nating by-products from the desired labelled compoundbefore HPLC purification [23,24] or by concentrating theradiopharmaceutical from HPLC solvent before formula-tion [24,25]. Liquid–liquid phase extraction [26] and RP-SPE [26–28] have also been used in metabolite studies toseparate unchanged tracer from radiolabelled metabolites.SPE is described as a rapid and reliable method [16]allowing the estimation of unchanged tracer in numeroussamples and is particularly advantageous in the determi-nation of the input function for compartmental modelfitting during PET investigations.

Our aim was to validate a rapid SPE method formetabolite analysis using HPLC and TLC in [18F]PBR102

and [18F]PBR111 rat and baboon studies, which would alsobe applicable to human studies.

2. Materials and methods

2.1. Radiotracers

The preparation of [18F]PBR102 and [18F]PBR111(Fig. 1) was performed using automated methods on theTracerlab FX FN and MX modules (GE Healthcare, UK)[29]. Briefly, 2–4 mg of the p-toluenesulfonyl precursorof PBR102 or PBR111, dissolved in anhydrous acetoni-trile, was added to pre-dried K[18F]–K2.2.2·K2CO3 com-plex, prepared from an aqueous [18F]fluoride solution(40–150 GBq). The reaction was heated at 100°C for 5 minbefore the reaction mixture was diluted with mobile phaseand purified by preparative reverse-phase chromatography.The collected radioactive peak was concentrated using aC18 SPE cartridge with subsequent formulation to aconcentration of 20 MBq/100 μl of saline containingb1% ethanol for the biological studies.

2.2. Animals and blood sampling

2.2.1. Rat studiesRat experiments were performed in compliance with the

NHMRC Australian Code of Practice for the care and use ofanimals for scientific purposes and were approved by theANSTO animal ethics committee.

Rats (Sprague Dawley) were anaesthetised with 5%isoflurane for induction and 2.5% for maintenance prior tothe procedures and metabolite studies. [18F]PBR radiotracers(30–40 MBq) were administered to 8- to 10-week-old malerats (200–250 g) via dorsal penile vein infusion (1 min) in avolume of 0.2 ml. Samples (nb11) of arterial blood (0.25 ml)were withdrawn via a cannula implanted in a femoral arteryfrom 1 to 120 min. Blood was centrifuged at 5000 rpm for5 min to isolate plasma. Plasma samples (0.1 ml) wereweighed and radioactivity was measured in a γ-counter.

2.2.2. Baboon studiesBaboon (Papio hamadryas) studies were approved by the

Sydney South West Area Health Service (SSWAHS) AnimalWelfare Committee.

139A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

PET-CT studies of [18F]PBR102 and [18F]PBR111 werecarried out for up to 2 h post isotope injection in malebaboons weighing 22.0±1.5 kg. The animals were studiedunder baseline conditions and at least 1 month later afterpreblocking or with displacement (injection time +30 min)using PK11195 (5 mg/kg) as competitor. Baboons wereanaesthetised with ketamine [6 mg/kg, given intramuscularly(im)] followed by a constant ketamine infusion in saline via aright antecubital vein. The baboons were also given atropine(0.04 mg/kg im) to prevent excess salivation, metoclopra-mide (5 mg im) to prevent vomiting and magnesium sulphate(1 g in 2 ml given slowly intravenously over 10 min) toprevent seizures. [18F]PBR radiotracers (150–200 MBq)were infused over a 3-min period through a left antecubitalvein. During the PET study, 16 arterial blood samples weretaken from a femoral artery from 0 to 120 min and processedas for the rat studies. Plasma samples (0.2 ml) were pipettedinto counting tubes and measured in a γ-counter.

2.2.3. Plasma sample preparation for metabolite analysisAfter counting, the determination of unchanged radio-

tracer in the plasma was performed by SPE and confirmed byradio-HPLC and radio-TLC analysis. For SPE and HPLC,100 μl (rat) or 200 μl (baboon) plasma was mixed with 200μl of water containing the unlabelled PBR compound (10nmol in 10 μl CH3OH) and KF (200 nmol) and analyseddirectly without protein elimination.

For radio-TLC analysis, plasma samples were treated asbefore and mixed with 0.5 ml acetonitrile. After proteinprecipitation and centrifugation (5000 rpm, 5 min), thesupernatant was evaporated under vacuum to dryness, re-dissolved in 25 μl methanol and applied to TLC plates. Theradioactivity of the precipitate was measured to quantify theacetonitrile extraction efficiency.

2.3. Solid-phase extraction

SPE analysis was performed on Oasis cartridges (WatersOasis HLB, 30 μm, 60 mg) that were preconditioned with 1ml ethanol followed by 5 ml water. Water (0.5 ml) was leftabove the sorbent until the use of the column.

To validate the extraction efficiency of the SPE methodand to determine the elution profile of the radioligands,non-radioactive rat and baboon plasma was mixed with[18F]PBR102 and [18F]PBR111 and with unlabelledPBR102 and PBR111 as described in the previous sectionand poured slowly onto the cartridge that was washedsuccessively with 1 ml of water, 1 ml of acetonitrile/water(20/80, v/v), 1 ml of acetonitrile/water (40/60, v/v), 1 mlof acetonitrile/water (60/40, v/v), 1 ml of acetonitrile/water(80/20, v/v) and 1 ml of acetonitrile.

In baboon PET experiments, after loading water-dilutedplasma samples from injected animals, each cartridge wassequentially washed with 1 ml of water, 1 ml of acetonitrile/water (20/80, v/v) and 2 ml of acetonitrile. Fractions werecollected into separate counting tubes. The radioactivity in

each fraction and that remaining on the cartridge weremeasured with a γ-counter.

2.4. High-performance liquid chromatography

Radio-HPLC analysis was performed on two systemsconsisting of a gradient pump (600, Waters), a 486 UVdetector set at 254 nm and an in-line detector formeasurement of radioactivity (Posiram IN/US) operatedby Waters Empower software using the same pre-column(Waters Oasis HLB, 25 μm, 3.9×20 mm) and two differentreverse-phase HPLC columns (System 1 for rat studies,Atlantis T3, 5 μm, 4.6×150 mm; System 2 for baboonstudies, Luna2 C18, 5 μm, 4.6×150 mm) according toliterature procedures [18,30]. In the first stage, afterloading samples, the Oasis pre-column was washed with1% acetonitrile in H2O for 3 min at 1.5 ml/min and thefraction containing protein residue and non-lipophiliccompounds by-passed the HPLC column while trappingnon-polar activity for subsequent separation. In the secondstage, the solvent direction was changed at 3 min to theHPLC column. Both columns in series in systems 1 and 2were then eluted with acetonitrile/water/TFA (35/65/0.1,v/v/v) as mobile phase at 1.5 ml/min flow rate for 20 min.The radioactivity peak corresponding to the authentic PBRcompounds was compared to the total activity registered inthe radiochromatogram to give the fraction of unchangedligand in the sample.

2.5. Thin layer chromatography

Ascending radio-TLC was performed on silica gel plateswith a pre-concentration zone (60 F254, Merck) using amixture of 20% methanol in ethylacetate. The distribution ofthe radioactivity was determined by phosphor plate imaging.TLC plates were exposed to Fujifilm Bas-Imaging plates MS4025 over 40 min. The exposed imaging plates wereevaluated using a Fujifilm Bio-imaging analyser BAS2500. After background correction, the fraction of un-changed [18F]PBR102 and [18F]PBR111 and that ofradioactive metabolites were calculated as the ratio ofphosphostimulated luminescence value in the identifiedregion using 254 nm UV light to the sum of all valuesmeasured in the chromatogram.

2.6. Validation of the chromatography techniques

For validation of the extraction efficiencies of [18F]PBR102 and [18F]PBR111 using the Oasis HLB cartridge inSPE analysis and pre-column in HPLC, 0.1–0.2 ml of non-radioactive plasma from rats or baboon was spiked with 20kBq of the radiofluorinated ligands. The activity that waswashed out with 2 ml of water from the SPE cartridge or with4.5 ml of 1% acetonitrile–water from the HPLC pre-column,according to the experimental conditions described above,was compared to the initial activity to determine the fractionof the radioligands not retained on the Oasis columns.

140 A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

Then the SPE fractions recovered from plasma samples ofanimals that had received [18F]PBR102 and [18F]PBR111were evaporated under vacuum to a minimum volume forHPLC analysis or to dryness and re-dissolved in methanolfor TLC analysis, to demonstrate that the SPE acetonitrilefraction contained pure [18F]PBR102 and [18F]PBR111 andthe absence of these [18F]PBR ligands in the other fractions.

2.7. Data analysis

The percentage of unchanged [18F]PBR102 and [18F]PBR111 in plasma of animals was fitted using GraphPadPrism software using the following bi-exponential equation:

y = y1 × e−k1� t� �

+ y2 × e−k2� t� �

where y1, y2 are the intercepts at t=0 of the two exponentialsand λ1, λ2 are the rate constants of the decays. Short andlong half-lives of the decay were then calculated according tot1/2=ln(2)/λ.

The plasma time–activity curves were corrected usingthe fraction of non-metabolised ligand measured by SPEfor each time point to determine the [18F]PBR102 and[18F]PBR111 input function curves for the compartmentmodel fitting of the tissue time–activity curves. [18F]PBR102 and [18F]PBR111 concentration in plasma wasnormalised for injected dose and expressed as % ofinjected dose per litre (%ID/L).

3. Results

3.1. Radiopharmaceutical preparation

The overall preparation of [18F]PBR102 and [18F]PBR111 including [18F]fluorination, HPLC purification,evaporation and radiotracer formulation was completedwithin 60 min with 25% radiochemical yield (non-decaycorrected) and radiochemical purity N99%. The specificactivity at injection time was N150 GBq/μmol.

3.2. Extraction efficiencies and SPE technique validation

Validation of the extraction process in SPE wasascertained using [18F]PBR ligands mixed with rat orbaboon plasma withdrawn prior to tracer injection. Morethan 99% of the activity remained on the SPE Oasis HLBcartridge after a three-step gradient elution from water to 40/60 acetonitrile/water. Moreover, when using 2 ml acetoni-trile directly after the 20/80 acetonitrile/water fraction (1 ml)in the SPE process, 93.0±2.4% (n=7) and 93.0±3.2% (n=6)of [18F]PBR102 and [18F]PBR111, respectively, were foundin the acetonitrile fraction, while b2% of the radioactivitywas still present on the SPE cartridge, indicating that therecovery from the cartridge was N95%.

In the HPLC analysis of [18F]PBR102 and [18F]PBR111standards mixed with animal plasma, no appreciableradioactivity was detected in the radiochromatogram withinthe first 3 min when the solvent was sent directly from the

precolumn to the radioactivity detector. Moreover, 93.8±3.1% and 96.3±2.4% of the radioactivity were found in the[18F]PBR102 or [18F]PBR111 peaks, respectively.

In the TLC analysis of [18F]PBR102 and [18F]PBR111standards mixed with plasma, and after plasma proteinprecipitation in the presence of carrier, the recovery ofradioactivity in the identified region of PBR compounds was94.8±2.9% (n=6) and 96.9±2.9% (n=8), respectively.Consequently, SPE, HPLC and TLC techniques wereefficient in the recovery of [18F]PBR102 and [18F]PBR111from plasma.

3.3. SPE analysis of plasma

Plasma samples of rats and baboons were analysed usingSPE up to 2 h after injection. Fig. 2 displays the typical timecourse of radioactivity distribution in each fraction afterinjection of [18F]PBR102 or [18F]PBR111 in animals. Inrats, initially, the percentage of the radioactivity in the pureacetonitrile fraction was high and diminished rapidly (Fig. 2,Panels A and B). Simultaneously, increasing percentages ofhydrophilic metabolites were seen in the aqueous washes,while lipophilic metabolites in 20% acetonitrile/waterfraction increased to a maximum between 5 and 15 min.

With the use of the same eluants as in rats, similar patternsof distribution of radioactivity from hydrophilic, lipophilicand unchanged compounds were observed in elutionfractions of the plasma analysis of baboons injected with[18F]PBR102 or [18F]PBR111 (Fig. 2, Panels C and D).

For the rat experiments, b3.5% of the radioactivity ofacetonitrile fraction remained on the Oasis cartridges.Similar results were obtained with the baboon plasmastudies; the activity remaining on Oasis cartridges wasassigned to the chemical species eluted by acetonitrile(unchanged tracer).

3.4. Plasma HPLC analysis

In HPLC radiochromatograms of plasma samples fromrats, three main radioactive peaks were observed. The frontpeak (0 to 3 min of retention time) corresponded tofragments having hydrophilic characteristics and not retainedby the Oasis HLB pre-column. The corresponding UV peakis related to plasma proteins. After switching to the HPLCcolumn, the radioactive peaks in the radiochromatogramappearing from 5 to 7 min were related to medium lipophilicradioactive metabolites analysed by the HPLC column andeluted rapidly with the 35% acetonitrile–water mixture. Inthese HPLC conditions, the retention times of the lipophilicparent compounds were 9 min for [18F]PBR102 and 11 minfor [18F]PBR111 as confirmed by the UV detection ofunlabelled standard.

In HPLC radiochromatograms of baboon plasma, as forrats, the solvent front corresponded to hydrophilic radio-metabolites not retained on the pre-column. After switchingthe solvent to the HPLC column, some radiometaboliteswere eluted rapidly from the HPLC column from 4 to 6 min.

Fig. 2. Typical time course of compositions of plasma radioactivity analysed using SPE, after injection of [18F]PBR102 (Panels A and C) or [18F]PBR111 (PanelsB and D) in rats under baseline conditions (Panels A and B) and in baboons under PK11195 pre-block conditions (Panels C and D) (expressed as % of totalplasma activity). Radioactivity in 2 ml acetonitrile fraction (□) and remaining in the Oasis cartridge (♢) represents unchanged radiotracers. Radioactivity in 2 mlwater fixation fraction (■) and in 1 ml (20/80) acetonitrile/water eluant (▲) represents hydrophilic and lipophilic radiometabolites, respectively. The fraction ofnon-metabolised [18F]PBR102 and [18F]PBR111 in the corresponding plasma sample, measured using radio-TLC in rats or using radio-HPLC in baboons, is alsorepresented (△).

141A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

The retention times of unchanged [18F]PBR102 and [18F]PBR111 were 16 to 17 min. Minor additional lipophilicpeaks were also observed eluting from the HPLC columnbefore the parent compound.

SPE analysis was used to confirm the absence ofunchanged tracers in the solvent front (0–3 min) that by-pass the HPLC column, indicating the entire recovery of[18F]PBR102 and [18F]PBR111 from plasma by the Oasispre-column.

3.5. Plasma TLC analysis

The same rat plasma samples analysed by HPLC werealso studied by radio-TLC. In the protein elimination processusing acetonitrile as denaturing agent, N92% of theradioactivity was recovered and further analysed by TLC.

Typical TLC radiochromatograms obtained for plasmasamples from a rat that received [18F]PBR102 are displayedin Fig. 3A. The region of interest at retention factor (Rf)=0.68corresponds to unchanged [18F]PBR102 as confirmed by themigration of [18F]PBR102 standard and the 254 nm UVabsorption of PBR102 added as carrier in the process. Noradioactivity was observed at RfN0.68, meaning that nometabolites more lipophilic than [18F]PBR102 were detectedusing the normal-phase TLC plate. Other minor spotscorresponding to lipophilic radioactive metabolites werenoticed at Rf=0.2–0.3. Quantification of the distribution ofthe radioactivity along each track correlated with the SPEand HPLC analysis.

Similar patterns of radioactivity distribution were ob-served in Fig. 3B for [18F]PBR111 with Rf=0.72 forunchanged tracer under the same TLC conditions.

Fig. 3. Phosphor plate images of the radio-TLC analysis of rat plasma samples (3, 5 and 30 min postinjection) and of fractions from SPE analysis (acetonitrile/water, 20/80, v/v and pure acetonitrile). Rats were injected with 30 MBq [18F]PBR102 (Panel A) and [18F]PBR111 (Panel B). The parent compounds areobserved at Rf=0.68 and 0.72, respectively.

142 A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

3.6. Validation of the chromatography techniques

3.6.1. TLC analysis of SPE fractionsThe efficiency of SPE to separate the parent compounds,

[18F]PBR102 or [18F]PBR111, from their radiolabelledmetabolites was demonstrated by analysing, after their γ-counting, the SPE fractions by radio-TLC and radio-HPLC.

The radiochromatograms of radio-TLC analysis of theacetonitrile/water (20/80, v/v) and acetonitrile SPE analysisfractions for 5min samples in comparison to the same plasmasample directly analysed by TLC are represented in Fig. 3.For [18F]PBR102 (Fig. 3, right side of Panel A), the spot(Rf=0.68) was only observed in the SPE acetonitrile fraction,while polar and lipophilic radioactive metabolites were onlynoticed in the acetonitrile/water (20/80, v/v) fraction. For[18F]PBR111, the same distribution of the radioactivityalong each track was also confirmed (Fig. 3, right side ofPanel B).

3.6.2. HPLC analysis of SPE fractionsThe radioactivity of the acetonitrile/water (20/80, v/v) and

acetonitrile SPE analysis fractions of 5-min plasma samplesfrom rats were also analysed using radio-HPLC (Fig. 4).Despite low counts in HPLC radiochromatograms, onlylipophilic metabolites were observed at 5 to 8 and 6 to 10minretention times in the SPE acetonitrile/water (20/80, v/v)fractions analysed (Panels A and C). On the contrary, SPEacetonitrile fractions contained N97% pure non-metabolised[18F]PBR102 (Panel B) and [18F]PBR111 (Panel D) asanalysed by HPLC.

3.7. Data analysis

The percentage of unchanged [18F]PBR102 or [18F]PBR111 in plasma of normal rats over time is plotted inFig. 5. Rapid decreases of the percentage of the parentcompounds were observed, and the time when 50% of intacttracers was present ranged from 15 to 20 min postinjection.After 60 min, the rate of decrease of the parent compoundswas slower and the fraction of unchanged [18F]PBR102(Panel A) was 15% of the total plasma radioactivity and 25%for [18F]PBR111 (Panel B).

Similar patterns were observed in baboon plasma asshown in Fig. 5 with two rates of decay of percentageunchanged [18F]PBR102 (Panel C) and [18F]PBR111 (PanelD). However, the time courses of the fraction of unchangedradiotracer in baboons were different between the baselineand blocking conditions. Initially, the % of unchanged tracerin plasma was clearly higher in animals that receivedPK11195 prior to the radiotracer.

The fractions of unchanged [18F]PBR determined usingSPE were compared to those obtained using HPLC, which isconsidered as the reference technique [16]. Fig. 6 comparesthe two chromatographic techniques in analysing the fractionof unchanged [18F]PBR102 and [18F]PBR111 in plasma ofbaboons under baseline and pre-blocking conditions byplotting the difference between the methods vs. the mean ofthe two methods (Bland–Altman plot [31]). For [18F]PBR102, the mean difference is close to 0 (0.9%), whilefor [18F]PBR111, a mean difference of 5.6% indicates ahigher estimate of percent non-metabolised tracer by the SPE

Fig. 4. HPLC radiochromatograms of SPE fractions of 5-min rat plasma injected with 30MBq of [18F]PBR102 (Panel A and B) or [18F]PBR111 (Panel C and D).After counting, the SPE fractions were evaporated to a minimum volume to eliminate acetonitrile and aliquots were then analysed using HPLC. Radiolabelledmetabolites are present in acetonitrile/water (20/80) fractions (Panels A and C), while acetonitrile fractions contained N97% pure [18F]PBR102 (Panel B) and[18F]PBR111 (Panel D).

143A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

method. The standard deviation (S.D.) of the difference isapproximately 5% for both tracers. There is a slight trend ofhigher non-metabolised fraction estimated with SPE (morepositive difference) with decreasing percent non-metabolisedtracer particularly for [18F]PBR111.

3.8. Curve fitting

In rats and in baboons, the curves expressing the % ofunchanged tracers were fitted using GraphPad software withbi-exponential decays (Fig. 5). The parameters of the curvefitting are reported in Table 1. In baseline and blockingstudies, the fraction of unchanged tracer determined by SPEfollowed a bi-exponential decrease with half-lives of 7 to 10min (fast component) and N80 min (slow component) for[18F]PBR102 and [18F]PBR111.

In baboons, the concentration of parent ligand in plasma(%ID/L) was derived from the radioactivity counting and thefraction of unchanged radiotracer was measured using SPE.

After a peak at 3–4 min, corresponding to the end of thetracer infusion, rapid decreases of unchanged [18F]PBR102and [18F]PBR111 concentration in plasma were observed(Fig. 7). The peak value radiotracer concentration was 3.7 to4.3 times higher in pre-blocking than in baseline (and beforethe displacement) experiments. Moreover, from the peak to30 min for baseline and displacement studies, the concen-tration in plasma of [18F]PBR102 and [18F]PBR111 wasreproducible and remained four times lower than in the pre-blocking condition.

In the displacement experiment, an increase in theconcentration of radiotracers in plasma was observedafter the intravenous injection of PK11195. This peakcorresponded to the displacement of [18F]PBR102 and[18F]PBR111 from peripheral binding sites. Then, after arapid decrease, the concentrations of [18F]PBR102 and[18F]PBR111 were similar to those of the pre-blockingexperiments and with the same slow component as baselinefor decay.

Fig. 5. Time course of the fraction of non-metabolised [18F]PBR102 and of [18F]PBR111 in plasma of rats (Panels A and B, respectively) and of baboons (PanelsC and D, respectively) measured using SPE. The fraction of non-metabolised tracer is expressed as a percent of total plasma activity (%). Rats were studied underbaseline conditions (n=4 for each radiotracer). Baboons received the radiotracer at baseline (♦) ([18F]PBR102: n=3; [18F]PBR111: n=4 ) and after pre-blocking(⋄) ([18F]PBR102: n=3; [18F]PBR111: n=2) with 5 mg/kg PK11195. Experimental data were fitted as the sum of two exponentials for baseline (solid line) andpre-blocking (dashed line) conditions.

144 A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

4. Discussion

We developed a rapid (extraction time about 15 s) andreliable method that uses Oasis HLB cartridges to extractquantitatively [18F]PBR102 and [18F]PBR111 from plasmafor subsequent separation using acetonitrile–water mixturesin SPE. The SPE analysis was validated with HPLC and TLCand can rapidly (sample can be processed in about 1 min)and, based on validation with HPLC and TLC, accuratelymeasure the unchanged [18F]PBR102 and [18F]PBR111fractions in plasma from rats and baboons. Our methodallows metabolite fractions to be determined for all plasmasamples and thus greatly facilitates the accurate estimation ofthe input function for measuring the receptor binding of these[18F]PBR tracers. The validation in two animal species,including a large primate, provides reassurance that thetechnique should be applicable to human studies.

4.1. Comparison of chromatography techniques

In preliminary experiments, the recovery of [18F]PBR102and [18F]PBR111 from standard solution mixed with plasmawas almost quantitative (N95%) from the SPE Oasis HLBcartridge. The quantitative extraction by SPE of [18F]PBR102 and [18F]PBR111 from plasma of rat or babooninjected with the radiotracers was ascertained by usingHPLC and TLC, which confirmed the absence of [18F]PBR102 and [18F]PBR111 in the water and acetonitrile/water (20/80, v/v) fractions.

The acetonitrile/water (20/80, v/v) solution was chosenbased on the elution profile of unchanged [18F]PBR102 and[18F]PBR111 in plasma, which showed that unchangedtracer was not eluted with mixtures containing b40%acetonitrile. The high lipophilicity value of PBR102 (logP7.5=2.7) and of PBR111 (log P7.5=3.2) measured by HPLC

Fig. 6. Bland–Altman plot comparing the fraction of non-metabolised [18F]PBR102 (Panel A) and [18F]PBR111 (Panel B) in plasma of baboon samples analysedusing SPE and HPLC. The mean difference and its standard deviation (±1 S.D.) are represented by a solid line and dashed lines, respectively. The fraction of non-metabolised [18F]PBR102 or [18F]PBR111, determined either by SPE or by HPLC, is the ratio of the concentration of unchanged [18F]PBR102 or [18F]PBR111to the total radioactivity concentration in plasma. The plasma samples were analysed using HPLC and SPE in parallel.

145A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

and reported by Fookes et al. [14] explains the strongadsorption on the hydrophilic–lipophilic balance sorbentused here. HPLC and TLC analysis of the elution profilefractions allowed the elution steps to be simplified to usingpure acetonitrile wash directly after the acetonitrile/water(20/80, v/v) elution. The acetonitrile fraction was confirmedto only contain unchanged tracer, irrespective of timepostinjection. This simplified procedure also successfullyachieved quantitative recovery of [18F]PBR ligands in theacetonitrile fraction. Therefore, the direct comparison of thedecay-corrected radioactivity in the acetonitrile wash plusactivity remaining on Oasis cartridges to that of the sum ofthe collected fractions represents the fraction of unchangedligands in the plasma. After decay correction, the radioac-tivity of the collected fractions was also compared to theactivity of the plasma before SPE analysis. The total activityof the plasma was recovered (100.7±3.0%, n=210) demon-

Table 1Parameters of the best fit of the curves expressing the fraction of unchangedtracers according to bi-exponential decays in rats and in baboons. The fittedcurves are shown in Fig. 5

y0=y1+y2(%)

y1(short)(%)

Shorthalf-life(min)

y2(long)(%)

Longhalf-life(min)

PBR102 Rat 72 55±1 10.1±3.7 17±5 145±350Baboonbaseline

70 42±7 7.6±2.7 28±4 185±190

Baboonblocking

91 56±9 7.0±2.1 35±6 120±80

PBR111 Rat 93 65±3 9.4±2.2 28±1 220±285Baboonbaseline

66 36±5 7.1±2.2 30±4 80±40

Baboonblocking

96 72±4 8.1±1.0 24±1 160±100

strating that there is no adsorption of lipophilic metabolitesor fluoride species on the glass tubes.

The quantitative extraction of PBR radiotracers by OasisHLB sorbent was also confirmed by HPLC analysis. Theactivity coming through the pre-column and by-passing theHPLC column was γ-counted, and no activity was found inthe first 4.5 ml 1% acetonitrile/water front when using[18F]PBR102 and [18F]PBR111 standards mixed withplasma. For plasma of injected animals, SPE and TLCanalysis of the initial 4.5 ml water front by-passing theHPLC column showed only hydrophilic metabolites and nointact [18F]PBR102 or [18F]PBR111. Therefore, the use ofOasis HLB-type columns allowed the total adsorption of[18F]PBR102 and [18F]PBR111 radiotracers from theplasma on the solid support, irrespective of whether theligands were free or bound to plasma proteins. Thequantitative extraction of all the [18F]PBR radiotracerswithout the need for prior plasma protein eliminationrepresents a definite advantage of the technique.

We assigned the fraction remaining on the SPE cartridgeto the chemical species eluted by acetonitrile (unchangedtracer) for the following reasons: the activity remaining onthe cartridge, expressed as fraction of activity eluted withacetonitrile, was independent of the blood sampling timeand, as a percentage of total activity, correlated linearly withthe fraction eluted with the 2-ml acetonitrile fraction(R2=0.65 to 0.90). If indeed a labelled metabolite wasretained on the SPE cartridge after the final elution, theradioactivity remaining on the cartridge for early bloodsamples would be low and, contrary to what we found, itwould not follow the high initial activity concentration of theunchanged tracer. Instead, the SPE cartridge fraction wouldbe expected to either increase continuously or present amaximum at some delayed time, similar to the metabolite

Fig. 7. Time course of plasma concentration of [18F]PBR102 (Panel A) and [18F]PBR111 (Panel B) after intravenous injection of [18F]PBR102 or [18F]PBR111in two baboons (weighing 22.4 and 21.4 kg, respectively) at baseline (⋄), pre-block (■) and displacement (△) with 5 mg/kg PK11195. The plasma activityconcentration was corrected using the fraction of non-metabolised ligand measured by SPE for each time point to determine the plasma concentration of intact[18F]PBR102 and [18F]PBR111. This concentration is expressed in %ID/L and displayed using a logarithm scale. The [18F]PBR111 baboon had twodisplacement studies; hence two displacement curves are shown, which are in good agreement.

146 A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

curve from the 20/80 acetonitrile/water fraction, which weagain did not observe. In addition, HPLC and TLC analysisdid not demonstrate any metabolites more lipophilic than theparent compounds. Furthermore, during the method devel-opment, using pure [18F]PBR102 and [18F]PBR111, the 2-ml acetonitrile volume dictated by gamma counter geometryconsiderations was insufficient to elute all the compound(b2% remaining on cartridge for 2 ml compared to b0.5%remaining for a 4 ml acetonitrile elution).

4.2. Comparison of biological results

In rat and in baboon, the time course of the unchangedfraction of [18F]PBR102 and [18F]PBR111 in plasma wassimilar with short (b10 min) and long (N80 min) half-livesfor the fast and slow components, respectively. Alkyl arylethers are known to dealkylate in mammals by a mono-oxygenase-mediated reaction, which occurs by hydroxyl-ation of the carbon alpha to the oxygen. The short half-lifebiotransformation of [18F]PBR102 and [18F]PBR111 inplasma and the rapid appearance of hydrophilic radiolabelledmetabolites were most likely due to [18F]fluoroacetate and3-[18F]fluoropropionate as the results of the dealkylation ofparent compounds [18F]PBR102 and [18F]PBR111 [14].The lipophilic metabolites are likely the result from othermetabolic pathways including N-dealkylation and furtherbiotransformation [32]. Similar rapid decreases in un-changed tracer have been observed in rat and baboon withthe structurally similar analogue [18F]-DPA714 [32,33] andwith other [18F]-PBR compounds, [18F]FEDAA [34], [18F]-FEPPA [35], [18F]PBR06 [36], having an F-18 atomattached on an aliphatic side chain.

As observed in Table 1, the sum (y0) of y1 (short) andy2 (long) values at the time of the beginning of theinjection is lower than 100%, ranging from 65% to 95%.

There are two possible explanations for these values: (i)[18F]PBR102 and [18F]PBR111 were infused over 1 minin rats and 3 min in baboon. During the infusion time, theconsequence of the rapid metabolism, corresponding to the8 to 10 min half-life, was to diminish the amount ofunchanged radioligand while the tracer was still beinginfused, particularly in baboon studies (3 min infusion).Therefore y1 (short) values were likely to be predominantlyaffected by the combination of rapid metabolism and slowinfusion. (ii) The avidity of PBRs in peripheral organs to[18F]PBR102 and [18F]PBR111 results in extraction of theligands from the plasma during infusion, leading to relativelyincreased fraction of radiolabelled metabolites. This hypoth-esis is supported by the pre-blocking studies. When the PBRswere occupied with PK11195 in pre-blocking experiments,they were unable to bind the radioligands that remained inblood. As consequence, the sums of unchanged y1 (shortspace)+y2 (long) in pre-blocking studies were higher thanthose in baseline experiments.

Our results demonstrate good agreement between SPEand HPLC techniques for [18F]PBR102 with a meandifference of percentage unchanged close to 0 and standarddeviation of 5.5%. For [18F]PBR111, the standard deviationof the difference was similar at 4.7%, but the mean differenceof 5.6% indicates higher estimates of the percent unchangedtracer with the SPE method, compared to HPLC. This isattributed to an underestimation of the radioactivity of thenon-metabolised radiotracer peak in comparison to the frontpeak in the HPLC chromatograms due to low counts in thispeak, particularly when the non-metabolised fraction issmall. This highlights the limitation of HPLC to accuratelydetermine percent unchanged tracer in low activity samples.SPE is a one-step method that allows the determination ofunchanged [18F]PBR compounds in samples with high andlow radioactivity concentrations from the plasma peak

147A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

concentration to the end of the study. Up to 16 plasmasamples were analysed over the 2h time course of the studyallowing accurate determination of [18F]PBR102 and [18F]PBR111 concentrations in plasma for input functiondetermination in PET. Interestingly, the time course ofunchanged radiotracer in plasma was reproducible inbaseline conditions and before the displacement withPK11195. The higher concentration of unchanged radio-tracers in plasma in blocking studies and after thedisplacement experiment with the competing PK11195 canbe explained by the reduced binding of [18F]PBR102 and[18F]PBR111 to peripheral binding sites under theseconditions. In displacement studies, PBR ligands bound toreceptors appear to have been protected from metabolism[14] and are then released into plasma. The same observa-tions have been made with [18F]PBR06 and [18F]FEPPAwhen using unlabelled drug or PK11195 in pre-blockingstudies in comparison to the concentration of unchangedligand in baseline conditions with high specific activityradiotracers [35,36].

5. Conclusion

Our method for direct plasma analysis of [18F]PBR102and [18F]PBR111 using RP-SPE is rapid and reliable. Theamount of unchanged tracers estimated by SPE stronglycorresponded with that determined by HPLC and TLCanalysis. SPE was superior to HPLC for low activity samplesand facilitates minimising the volume of withdrawn bloodsamples. SPE allowed rapid and accurate metaboliteanalysis, thus overcoming the limitation of HPLC in onlyallowing metabolite analysis for a limited number of plasmasamples used to derive the input function. In this study, weshowed that speed of SPE analysis allowed metaboliteanalysis of all input function samples, thus improving theaccuracy of the metabolite-corrected input function forquantitative imaging studies using compartmental models.The ability to analyse every plasma sample with SPE methodalso greatly facilitates characterisation of the non-monotonicintact tracer fraction time course for displacement studies.

References

[1] Anholt RR, De Souza EB, Oster-Granite ML, Snyder SH. Peripheral-type benzodiazepine receptors: autoradiographic localization in whole-body sections of neonatal rats. J Pharmacol Exp Ther 1985;233:517–26.

[2] Papadopoulos V, Amri H, Boujrad N, Cascio C, Culty M, Garnier M,et al. Peripheral benzodiazepine receptor in cholesterol transport andsteroidogenesis. Steroids 1997;62:21–8.

[3] Banati RB. Visualising microglial activation in vivo. Glia 2002;40:206–17.

[4] Venneti S, Bonneh-Barkay D, Lopresti BJ, Bissel SJ, Wang G, MathisCA, et al. Longitudinal in vivo positron emission tomography imagingof infected and activated brain macrophages in a macaque model ofhuman immunodeficiency virus encephalitis correlates with central andperipheral markers of encephalitis and areas of synaptic degeneration.Am J Pathol 2008;172:1603–16.

[5] Pike VW, Halldin C, Crouzel C, Barre L, Nutt DJ, Osman S, et al.Radioligands for PET studies of central benzodiazepine receptors andPK (peripheral benzodiazepine) binding sites-current status. Nucl MedBiol 1993;20:503–25.

[6] Charbonneau P, Syrota A, Crouzel C, Valois JM, Prenant C, Crouzel M.Peripheral-type benzodiazepine receptors in the living heart character-ized by positron emission tomography. Circulation 1986;73:476–83.

[7] Pappata S, Cornu P, Samson Y, Prenant C, Benavides J, Scatton B,et al. PET study of carbon-11-PK11195 binding to peripheral typebenzodiazepine sites in glioblastoma: a case report. J Nucl Med1991;32:1608–10.

[8] Banati RB, Newcombe J, Gunn RN, Cagnin A, Turkheimer F, HeppnerF, et al. The peripheral benzodiazepine binding site in the brain inmultiple sclerosis — quantitative in vivo imaging of microglia as ameasure of disease activity. Brain 2000;123:2321–37.

[9] Vowinckel E, Reutens D, Becher B, Verge G, Evans A, Owens T, et al.PK11195 binding to the peripheral benzodiazepine receptor as a markerof microglia activation in multiple sclerosis and experimental autoim-mune encephalomyelitis. J Neurosciences Res 1997;50:345–53.

[10] Cagnin A, Gerhard A, Banati RB. In vivo imaging of neuroinflamma-tion. Eur Neuropsychopharmacology 2002;12:581–6.

[11] Gerhard A, Banati RB, Goerres GB, Cagnin A, Myers R, Gunn RN,et al. [C-11](R)-PK11195 PET imaging of microglial activation inmultiple system atrophy. Neurology 2003;61:686–9.

[12] Chauveau F, Boutin H, Van Camp N, Dolle F, Tavitian B. Nuclearimaging of neuroinflammation: a comprehensive review of [C-11]PK11195 challengers. Eur J Nucl Med Mol Imaging 2008;35:2304–19.

[13] Dolle F, Luus C, Reynolds A, Kassiou M. Radiolabelled molecules forimaging the translocator protein (18 kDa) using positron emissiontomography. Curr Med Chem 2009;16:2899–923.

[14] Fookes CJ, Pham TQ, Mattner F, Greguric I, Loc'h C, Liu X, et al.Synthesis and biological evaluation of substituted [18F]imidazo[1,2-a]pyridines and [18F]pyrazolo[1,5-a]pyrimidines for the study of theperipheral benzodiazepine receptor using positron emission tomogra-phy. J Med Chem 2008;51:3700–12.

[15] VanCampN,BoigardR,Kuhnast B,ThezeB,Viel T,GregoireMC, et al.In vivo imaging of neuroinflammation; a comparative study between[18F]PBR111, [11C]CLINME and [11C]PK11195 in an acute rodentmodel. Eur J Nucl Med Mol Imaging 2010;37:962–72.

[16] Maziere B, Cantineau R, Coenen HH, Guillaume M, Halldin C, LuxenA, et al. PET radiopharmaceuticals metabolism — plasma metaboliteanalysis. In: Stocklin G, Pike VW, editors. Radiopharmaceuticals forpositron emission tomography. Dordrecht: Kluwer Academic Publish-ers; 1993, pp. 151–78.

[17] Van Waarde A, Visser TJ, Posthumus H, Elsinga PH, Anthonio RL,Van Loenen-Weemaes AMA, et al. Quantification of the beta-adrenoceptor ligand S-1′-[18F]fluorocarazolol in plasma of humans,rats and sheep. J Chromato B 1996;678:253–60.

[18] Hilton J, Yokoi F, Dannals RF, Ravert HT, Szabo Z, Wong DF.Column-switching HPLC for the analysis of plasma in PET imagingstudies. Nucl Med Biol 2000;27:627–30.

[19] Luthra SK, Osman S, Turton DR, Vaja V, Dowsett K, Brady F. Anautomated system based on solid phase extraction and HPLC for theroutine determination of unchanged [11C]diprenorphine, [11C]fluma-zenil, [11C]raclopride and [11C]Scherring 23390. J Label CompdRadiopharm 1993;32(S1):518–20.

[20] Greuter H, Van Ophemert PLB, Luurtsema G, van Berckel BNM,Franssen EJF, Windhorst BD, et al. Optimizing an online SPE–HPLC method for analysis of (r)-[C-11]1-(2-chlorophenyl)-n-methyl-n-(1-methylpropyl)-3-isoquinolinecarboxamide [(r)-[C-11]PK11195] and its metabolites in humans. Nucl Med Biol 2005;32:307–12.

[21] Chitneni SK, Serdons K, Evens N, Fonge H, Celen S, Deroose CM, et al.Efficient purification and metabolite analysis of radiotracers using high-performance liquid chromatography and on-line solid-phase extraction.J Chrom A 2008;1189:323–31.

148 A. Katsifis et al. / Nuclear Medicine and Biology 38 (2011) 137–148

[22] Gillings N. A restricted access material for rapid analysis of [11C]-labeled radiopharmaceuticals and their metabolites in plasma. NuclMed Biol 2009;6:961–5.

[23] Kiesewetter DO, Kilbourn MR, Landvatter SW, Heiman DF,Katzenellenbogen JA, Welch MJ. Preparation of four fluorine-18-labeled estrogens and their selective uptakes in target tissues ofimmature rats. J Nuc Med 1984;25:1212–21.

[24] Kiesewetter DO, Lee JT, Lang L, Park SH, Paik CH, Eckelman WC.Preparation of 18F-labeled muscarinic agonist with M2 selectivity.J Med Chem 1995;38:5–8.

[25] Lemaire C, Plenevaux A, Aerts J, Del Fiore G, Brihaye C, Le Bars D,et al. Solid phase extraction — An alternative to use of rotaryevaporators for solvent removal in the rapid formulation of PETradiopharmaceuticals. J Label Compd Radiopharm 1999;42:63–75.

[26] Ma Y, Kiesewetter DO, Lang L, Der M, Huang B, et al. Determinationof [18F]FCWAY, [18F]FP-TZTP, and their metabolites in plasma usingrapid and efficient liquid–liquid and solid phase extractions. Nucl MedBiol 2003;30:233–40.

[27] Muzic RF, Landmeier B, Zhang Z, Zheng L, Berridge MS. Solid-phaseanalysis method for (S)-[18F]fluorocarazolol and its metabolites.J Chromato B 2001;759:355–9.

[28] Shumway DA, Pavlova OA, Mukhin AG. A simplified method for themeasurement of nonmetabolised 2-[18F]FA-85380 in blood plasmausing solid-phase extraction. Nucl Med Biol 2007;34:221–8.

[29] Pham T, Bourdier T, Henderson D, Fookes C, Greguric I, Jackson T, etal. Automated radiosynthesis of [18F]PBR111 and [18F]PBR102-selective radioligands for imaging the peripheral benzodiazepinereceptor with PET. J Label Compd Radiopharm 2009;52(S1):310.

[30] Wilson AA, McCormick P, Kapur S, Willeit M, Garcia A, HusseyD, et al. Radiosynthesis and evaluation of [11C]-(+)-4-propyl-3,4,4a,5,6,10b-hexahydro-2H-naphtho[1,2-b][1,4]oxazin-9-Ol as apotential radiotracer for in vivo imaging of the dopamine D2 high-affinity state with positron emission tomography. J Med Chem2005;48:4153–60.

[31] Bland JM, Altman DG. Statistical methods for assessing agreementbetween two methods of clinical measurement. Lancet 1986;1:307–10.

[32] Peyronneau M, Damont A, Valette H, Saba W, Delforge J, Goutal S, etal. Metabolism of DPA-714, a new peripheral benzodiazepine receptorPET ligand. J Label Compd Radiopharm 2009;52(S1):385.

[33] Chauveau F, Van Camp N, Dolle F, Kuhnast B, Hinnen F, Damont A,et al. Comparative evaluation of the translocator protein radioligands11C-DPA-713, 18F-DPA-714, and 11C-PK11195 in a rat model of acuteneuroinflammation. J Nucl Med 2009;50:468–76.

[34] Zhang MR, Maeda J, Ogawa M, Noguchi J, Ito T, Yoshida Y, et al.Development of a new radioligand, N-(5-fluoro-2-phenoxyphenyl)-N-(2-[18F]fluoroethyl-5-methoxybenzyl)acetamide, for PET imaging ofperipheral benzodiazepine receptor in primate brain. J Med Chem2004;47:2228–35.

[35] Wilson AA, Garcia A, Parkes J, McCormick P, Stephenson KA, HouleS, et al. Radiosynthesis and initial evaluation of [18F]-FEPPA for PETimaging of peripheral benzodiazepine receptors. Nucl Med Biol2008;35:305–14.

[36] Briard E, Zoghbi SS, Simeon FG, ImaizumiM, Gourley JP, Shetty HU,et al. Single-step high-yield radiosynthesis and evaluation of asensitive, F-18-labeled ligand for imaging brain peripheral benzodi-azepine receptors with PET. J Med Chem 2009;52:688–99.