Effect of Reduction in EndogenousDopamine on Extrastriatal Binding of

[11C]FLB 457 in Rat Brain—An Ex Vivo Study

RABIA AHMAD,1* ELLA HIRANI,1 PAUL M. GRASBY,2 AND SUSAN P. HUME1

1Hammersmith Imanet Ltd., Hammersmith Hospital, Du Cane Road, London W12 0NN, United Kingdom2MRC Clinical Sciences Centre, Cyclotron Building, Hammersmith Hospital, London, W12 0NN, United Kingdom

KEY WORDS FLB 457; dopamine D2 receptor; D3 receptor; gamma-butyrolactone;radioligand; rat

ABSTRACT Carbon-11 labeled FLB 457 has been used successfully as a selective,high affinity PET ligand for the quantification of extrastriatal D2-like receptors inman. This study was carried out in rats to investigate regional values for maximalbinding and ED50 (a measure of apparent Kd) for the radioligand in vivo in controlanimals and in a group pretreated with the neuronal impulse flow inhibitor, g-butyro-lactone. The aims were to obtain further information regarding the specific activityneeded to ensure tracer kinetics and to investigate baseline occupancy by dopamine(DA), each relevant to optimal clinical use of the radioligand. Regional Bmax valueswere consistent with the distribution of D2-like receptors in rat brain. Of interest,60% of the binding in cerebellum, often used as a low-binding ‘‘reference region’’ forPET quantification, was saturable, with Bmax only 2- to 3-fold less than that in neocor-tex, hippocampus, and thalamus. ED50 values were in the range 2–3 nmol/kg, confirm-ing minimal receptor occupancy by the tracer in human PET, using high but achiev-able specific activities. In the majority of extrastriatal tissues, reduction in synapticDA did not significantly decrease the apparent Kd, except in cortical regions, wherethe extent of the effect suggested a low (�10%), but measurable baseline receptoroccupancy by DA. Synapse 59:162–172, 2006. VVC 2005 Wiley-Liss, Inc.

INTRODUCTION

The use of positron emission tomography (PET) forin vivo studies of dopaminergic dysfunction is welldeveloped, with radiotracers established for monitor-ing both pre- and postsynaptic dopamine (DA) bindingsites, primarily in the striata. The radioligand mostfrequently used to quantify DA D2-like receptors is[11C]raclopride (Farde et al., 1986), with specific bind-ing usually estimated using a reference region (cerebel-lum) modeling approach (Lammertsma et al., 1996).More recently, alternative D2 receptor radioligandshave been developed as the use of PET, together withsingle photon emission tomography (SPET), hasexpanded (de Paulis, 2003). For example, the highaffinity ligand, [11C]FLB 457 (Halldin et al., 1995) hasbeen used to measure D2-like receptors in extrastriataltissues, where their density (Bmax) is an order of mag-nitude lower than in the striatum (Camps et al., 1990;Kessler et al., 1993).

The D2/D3 receptor selective compound, FLB 457((S)-N-((1-ethyl-2-pyrrolidinyl)methyl)-5-bromo-2,3-di-

methoxybenzamide), or isoremoxipride, is one of a se-ries of substituted benzamides, designed to improve theantipsychotic effect of atypical neuroleptics, with lowerincidence of extrapyramidal side effects (Hogberg,1991). Compared with an in vitro dissociation constant(Kd) of �1.4 nM for radiolabeled raclopride (Kohleret al., 1985), the in vitro inhibition constant (Ki) of FLB457 has been reported as �0.02 nM (Halldin et al.,1995). Since in vivo binding potential (BP) approxi-mates to Bmax/Kd (Farde et al., 1989), this high affinityproperty, taken alongside low nonspecific binding,makes [11C]FLB 457 a potentially more appropriatePET ligand than [11C]raclopride for quantifying extra-striatal DA receptors, although its slow kinetics negate

*Correspondence to: Rabia Ahmad, Hammersmith Imanet Ltd, Hammer-smith Hospital, Du Cane Road, London W12 ONN, United Kingdom.E-mail: rabia.ahmad@imperial.ac.uk

Received 6 July 2005; Accepted 4 October 2005

DOI 10.1002/syn.20231

Published online in Wiley InterScience (www.interscience.wiley.com).

VVC 2005 WILEY-LISS, INC.

SYNAPSE 59:162–172 (2006)

its use for accurate estimation of striatal BP (Olssonand Farde, 2001).

High affinity ligands do, however, require high spe-cific activities to satisfy ‘‘tracer’’ kinetics (Sandell et al.,2000). For human PET, Sudo et al. (2001), for example,have suggested a coinjected FLB 457 dose of <1.6nmol/body to limit mass effects to below 5%. The cur-rent study, measuring in vivo binding of [11C]FLB 457in rats, used the approach outlined by Hume et al.(1997, 1998) to determine regional in vivo ED50, thedose required to saturate half of the specific bindingand, thus, related to the apparent Kd. The aim was tobetter estimate the specific activity required for mini-mal receptor occupancy in experimental and (feasibly)human PET. Since FLB 457 has a similar affinity forthe D2 and D3 receptor subtypes (Halldin et al., 1995)and D3 receptors are present in the cerebellum (Ban-croft et al., 1998; Levant, 1997; Vessotskie et al. 1997),our study included this low binding region.

In addition, we studied the effect of a reduction inDA concentration on extrastriatal binding to determinebaseline DA occupancy. The theoretical approach wasas documented for striatal D2 receptors (Laruelle,2000). Assuming a simple competition between radioli-gand and neurotransmitter, the apparent Kd of[11C]FLB 457 may include a component of occupancydue to DA, dependent on the latter’s concentration andKi. Thus, specific binding of the radioligand can bemoderated by competition with the neurotransmitter,according to the following relationship: BP ¼ Bmax/(ligand Kd*(1 þ [DA]/Ki)), as detailed in Ginovart et al.(1997). Since DA has a high affinity for D3 receptors(e.g., Levesque et al., 1992), in vivo occupancy or occlu-sion by neurotransmitter might be higher than thatexpected for D2 receptors (Schotte et al., 1992). Wehypothesized a reduction in ED50 for [11C]FLB 457 fol-lowing DA depletion, with the possibility of regionaldifferences, dependent not only on DA concentrationbut also on the relative distributions of the D2 and D3

subtypes. The findings are pertinent to interpretationof [11C]FLB 457 binding changes in patient populationswhere altered levels of DA might be expected (Talviket al., 2003) and, additionally, could provide experimen-tal evidence for the use of PET with [11C]FLB 457 tomonitor acute changes in DA concentration.

Depletion of synaptic DA was achieved by pretreat-ing the rats with g-butyrolactone (GBL), which causesinhibition of impulse flow in both nigrostriatal andmesolimbocortical neurons (Maitre, 1997; Wolf andRoth, 1987). In an earlier microdialysis/PET study inrats (Hume et al., 1992), we reported a rapid and pro-longed �75% reduction in extracellular striatal DA fol-lowing the dose of GBL used in the present study. Here,we used ex vivo dissection rather than small animalPET to avoid both partial volume effects and the com-plication of anesthesia.

MATERIALS AND METHODS

All work was carried out by licensed investigators inaccordance with the UK Home Office’s ‘‘Guidance onthe Operation of the Animals (Scientific Procedures)Act 1986’’ (HMSO, Feb 1990) and used adult maleSprague–Dawley rats (mean 6 SE body weight ¼ 2986 4 (n ¼ 61)). Animals were purchased from HarlanUK Ltd., Bicester, Oxon, UK. GBL was from Sigma-Aldrich, Poole, Dorset, UK and FLB 457 from ABX,Radeberg, Germany. Carbon-11 labeled FLB 457 wassynthesized for clinical studies by O-methylation of itsdesmethyl precursor using [11C]methyl triflate(adapted from Sandell et al., 2000), and aliquots takenfor biological work.

Kinetic study

Since it was considered not feasible to obtain fullradioactivity:time profiles for the many doses of FLB457 needed for the saturation studies, a preliminarykinetic study was carried out to select an optimal sam-pling time. The radioligand (mean 6 SD radiochemicalpurity of (97 6 2)%) was given i.v. to awake but lightlyrestrained rats (n ¼ 14), via a previously catheterizedtail vein. The mean 6 SD injectate was 11.7 60.9 MBq, with a stable FLB 457 content of 0.7 6 0.3nmol/kg. Arterial blood was collected at graded times(5 samples/rat) via a previously catheterized tail arteryand, in selected samples, the proportion of radiolabeledmetabolites in plasma was measured by reverse phasehigh performance liquid chromatography. Previously,others have shown no measurable radiolabeled metab-olites in rat brain (Suhara et al., 1999).

At designated times, between 1 and 120 min after[11C]FLB 457 administration, rats were killed and thebrains rapidly removed. Twelve regions were sampled,as listed in Table I, and were chosen to represent tis-sues with a range of D2-like receptor density. Radioac-tivity was measured using a Wallac gamma-counter,

TABLE I. Regional binding potential

Tissue BP

Olfactory tubercles 5.9 6 1.4Hypothalamus 3.3 6 0.9Thalamus 2.5 6 0.6Prefrontal cortex 2.1 6 0.3Striatum (�83)Frontal with parietal cortex 2.3 6 0.5Hippocampus 1.8 6 0.2Occipital with temporal cortex 1.1 6 0.2Inferior colliculi 11 6 9Superior colliculi 5.1 6 2.8Pons with Medulla 2.7 6 0.5Cerebellum 0.82 6 0.09

Regional binding potential (6SE) estimated from best fits of radioactivity:timedata (1–120 min). Data were fitted to a two tissue compartmental model, usinga metabolite corrected plasma input. The number annotation corresponds withthat used subsequently in Figures 2, 5, and 7. The striatal BP (shown inbrackets) had a high associated error.

163EXTRASTRIATAL BINDING OF [11C]FLB 457

with automatic correction for radioactive decay, andresults normalized for the amount injected relative tobody weight, giving

uptake units ¼ (cpm/g wet weight tissue)=

(injected cpm/g body weight)

Plasma radioactivity concentration data were fitted toa multiexponential function and corrected for radiola-beled metabolites. Tissue radioactivity concentrationdata were then fitted to a two-tissue compartmentalmodel using the plasma parent profile as input toobtain estimates of BP, the ratio of k3/k4, or the spe-cific-to-nonspecific partition coefficient (Lammertsmaet al., 1996). Animal numbers were kept to a mini-mum by using only one rat per time point, but with asufficient number of times to allow satisfactory ki-netic analysis of the data. Of relevance to the ex-pected error on each time point, the SD on regionaluptake values from three rats killed at 60 min after[11C]FLB 457 (coinjected stable 0.6–0.7 nmol/kg) was�14% of the mean (see below).

To check that the time course of development of sig-nal was not grossly changed by administration of GBL,a further group of rats (n ¼ 11) were given GBL at ani.p. dose of 400 mg/kg, 40 min before [11C]FLB 457(10.0 6 0.2 MBq; stable FLB 457 ¼ 0.4 6 0.2 nmol/kg).Sampling was as above, again with one rat per timepoint (graded between 2 and 90 min after radioligandinjection), except that plasma was obtained postmor-tem and was not analyzed for radioactive metabolites.Those rats treated with GBL appeared sedated andremained so throughout the assay period.

Saturation studies

Two groups of rats were used: control (n ¼ 18, includ-ing 1 from the first study) and a group pretreated withGBL, administered as above (n ¼ 20, including 1 fromthe first study). The mean 6 SE specific activities ofthe radiosynthesized compound were 64 6 5 MBq/nmolfor the control group (no placebo injection) and 71 6 12MBq/nmol for the GBL-treated group, decay-correctedto the first injection for each experiment. Saturation ofbinding sites was determined by varying the mass ofcompound injected, achieved either by dilution of injec-tate or by predosing rats with FLB 457, given i.v.,5 min before the radioligand. The final injectate for thecontrol group ranged between 1.4–21.5 MBq and0.042–2696 nmol/kg, with equivalent ranges of 0.5–11.2 MBq and 0.043–2696 nmol/kg for the GBL-treatedgroup. The minimal FLB 457 dose was limited by theamount of radioactivity needed to achieve adequatecounts. In both series of studies, sample data were dis-carded if the error on the radioactive counts wasgreater than �5%.

Rats were euthanased 60 min after radioligandadministration and regions sampled as describedabove. Uptake data were normalized to individualplasma con- centration (obtained immediately postmortem) with the aim of reducing inter-rat variability.Using the 3-rat example cited above, the average coeffi-cient of variation across regions decreased from 14 68% to 11 6 3%. Tissue:plasma data were thenexpressed as a function of FLB 457 dose. If it is consid-ered that an equilibrium exists between plasma levelsof radioligand and its concentration at the site ofaction, then using the Michaelis–Menten relationshipand a single competition/binding site, the concentra-tion of specifically bound ligand (RL) is related to Bmax,apparent Kd (or ED50), and radioligand dose (L) as

RL ¼ ðBmax3LÞ=ðED50 þ LÞ

The parameters Bmax, ED50, and NS (the nonspecificcomponent that cannot be eliminated by receptor sat-uration) were then obtained by iterative nonlinearregression, for each region. ED50 is expressed in unitsof injected FLB 457 (i.e., nmol/kg). Full details of therationale and parameter derivation are given inHume et al. (1997). Best-fits to the two-site competi-tive binding model gave unacceptably wide confidenceintervals for the fitted parameters.

Analyses

In-house software packages were used for all modelfitting, with equal weighting given to each point (Wal-lac count rate error on each sample < 5%). Statisticalanalyses (primarily two-way ANOVA with Bonferronicorrection for multiple comparisons) and comparison ofthe kinetic model fits (corrected Akaike’s InformationCriteria and F test) were carried out using GraphPadPrism version 4.00, GraphPad Software, San Diego,California, USA.

RESULTS

Following i.v. [11C]FLB 457, radioactivity was rap-idly and highly extracted into all brain regions dis-sected, giving uptake units of the order of 2 at the firstsampling time. Thereafter, radioactivity was furtheraccumulated only in striata. All remaining regionsshowed a loss of radioactivity, the rate of which was tis-sue dependent. Figure 1 illustrates the radioactivityconcentration:time profiles for three of the regions,namely hypothalamus, prefrontal cortex, and cerebel-lum, representing extrastriatal tissues with moderate,low, and minimal D2-like receptor density. The solidlines represent the best fit to the two-tissue compart-ment model, using a metabolite-corrected plasma inputfunction (dotted line). BP estimates for the illustratedregions are given in the figure and all values are listedin Table I. In each instance, including cerebellum, an

164 R. AHMAD ET AL.

unsatisfactory fit was obtained using a single-tissuecompartment model (fits compared using correctedAkaike’s Information Criteria and F test models, datanot shown).

Figure 2 shows the relationship between extrastria-tal BP and radioactivity concentration measured at60 min, a time chosen as a compromise between thatneeded to reach equilibrium and that to ensureadequate count statistics in small region of interest(ROI). Although the shortened sample time resulted ingross underestimation of binding capacity in striata,there was a good linear correlation (R ¼ 0.952, P <0.001, by least squares regression) for the otherregions, including olfactory tubercles, which might beconsidered as part of the striatal complex. In Figure 2,the 60-min data were estimated from best fits such asthose illustrated in Figure 1 but a similar correlation(R ¼ 0.953) was also obtained for the measured data.

Since GBL altered the behavior of the rats and mightbe expected to change, for example, local perfusionand/or metabolism (Wolfson et al., 1977), the adequacyof a 60-min time point was also checked in a GBL-treated group of rats. Full kinetic modeling of the datawas not possible as, due to a resource limitation,plasma counts were not corrected for radiolabeledmetabolites. As a compromise, cerebellum was ac-cepted as approximating to a reference region, with amajor part of its radioactivity concentration expectedto reflect nonspecific binding. Figure 3 illustrates thedevelopment of the in vivo signal as a function of timein inferior colliculi, a relatively high binding region.

Fig. 1. Radioactivity content in three of the brain regionssampled, as a function of time after i.v. injection of [11C]FLB 457:(a) hypothalamus (b) prefrontal cortex and (c) cerebellum. Eachdatum point is from an individual rat, with left and right-sidedregions pooled. The solid lines are the fits to a two-tissue compart-ment model using metabolite-corrected plasma radioactivity concen-tration (dotted line) as the input function.

Fig. 2. Regional radioactivity content (uptake units) at 60 minafter injection of [11C]FLB 457 as a function of BP, estimated fromthe best fits to the dynamic data acquired over 120 min. Eachdatum point represents a discrete region, as listed in Table I. Thefitted line is the linear correlation obtained by least squares (R ¼0.952, P < 0.001) for the 11 extrastriatal tissues.

165EXTRASTRIATAL BINDING OF [11C]FLB 457

Here, and in all other regions sampled, there was nogross effect of GBL on the rate of development of spe-cific binding, expressed as either (a) tissue:cerebellumratio or (b) tissue:total plasma count ratio, validatingthe 60-min time point as a measure of BP for bothgroups. Based on these findings, 60-min data normal-ized to individual rat plasma (see Methods) were usedas a surrogate measure of BP in the following majorpart of the study designed to determine the regionalsaturability of extrastriatal [11C]FLB 457 binding.

Fig. 3. Time course of development of ‘‘specific signal’’ in inferiorcolliculi of rats pretreated with GBL (open symbols) compared withcontrol (closed symbols), showing tissue radioactivity concentrationas a function of time, expressed relative to either (a) individual cer-ebellum concentration or (b) plasma concentration, the latter uncor-rected for radiolabeled metabolites. The data are from single ratsper time point; the control series was that used to obtain the BPvalue presented in Table I.

Fig. 4. Tissue concentration of [11C]FLB 457 in control rats,60 min after radioligand injection, as a function of FLB 457 dose,for (a) hypothalamus (b) prefrontal cortex and (c) cerebellum. Eachdatum point is from an individual rat, normalized to plasma radio-activity. The solid lines represent the best fits to a single-site bind-ing model, as discussed in the Materials and Methods section andexplained in detail in Hume et al. (1997).

166 R. AHMAD ET AL.

Examples of the saturation curves obtained in con-trol rats are illustrated in Figure 4, for the threeregions shown in Figure 1. In each, including cerebel-lum, there was a dose-dependent decrease in 60-mintissue:plasma ratio, to a minimum value. ED50 esti-mates for the illustrated regions are given in the fig-ure. Estimates for Bmax, ED50, and NS for all regionsare given in Table II. With the exception of olfactorytubercles, where the ED50 value was �10 nmol/kg,ED50 values were in the range 1.9–3.7 nmol/kg. Bmax

values ranged from 63 nmol/kg in olfactory tubercles to2.4 nmol/kg in cerebellum. The nonspecific componentwas similar for all regions, with a tissue:plasma ratioof the order of 0.6. The fact that this is less than unitypresumably reflects, at least in part, the contributionof radiolabeled metabolites to plasma radioactivity.

The maximal binding ratios (at ‘‘trace’’ levels ofradioligand) for each region were estimated as (Bmax/ED50) þ NS and are listed in Table II. At an injecteddose of �0.7 nmol/kg, as previously used for the kineticstudy in control rats, measurable occupancy would beexpected for all extrastriatal regions, with the excep-tion of olfactory tubercles. Although the rank order ofin vivo binding remained the same, the resultant effect(determined retrospectively) was an underestimationof BP, by �20% using a mean ED50 value of 2.7 nmol/kg.Figure 5 illustrates the correlation between the esti-mated maximal signal (tissue:plasma ratio at 60 min)from the saturation study and the measured BP fromthe kinetic study.

Saturation curves from GBL-treated rats are shownin Figure 6, for the same regions as shown for controlrats in Figure 4. The estimates for Bmax, ED50, and NSfor all regions are listed in parentheses in Table II.

Two-factor analysis of variance showed no statisticallysignificant difference in NS between groups or acrosstissues. Although in many cases, the GBL-treated ratsshowed similar saturation kinetics to the controlgroup, there were some statistically significant differ-ences in Bmax and ED50, which, as predicted, wereregion-dependent. Thus, while analysis of variance ofBmax showed no overall effect of GBL (P ¼ 0.07), a

TABLE II. Saturation parameters

Tissue Bmax ED50 NS Maximal ratio Occupancy

Olfactory tubercles 63.1 6 24.9 10.5 6 4.0 0.4 6 0.4 6.41(42.7 6 21.2) (7.3 6 3.5) (0.8 6 0.6) (6.65) þ4%

Hypothalamus 11.5 6 2.6 1.9 6 0.5 0.7 6 0.2 6.75(11.4 6 2.7) (1.9 6 0.5) (0.6 6 0.2) (6.60) �2%

Thalamus 11.5 6 3.5 3.3 6 1.0 0.6 6 0.2 4.08(11.7 6 3.7) (2.9 6 0.9) (0.4 6 0.2) (4.43) þ8%

Prefrontal cortex 7.4 6 2.6 2.8 6 1.0 0.6 6 0.1 3.24(3.9 6 1.2) (1.3 6 0.5) (0.6 6 0.1) (3.60) þ10%

Frontal with Parietal cortex 8.3 6 4.7 3.7 6 2.2 0.6 6 0.2 2.84(4.0 6 1.7) (1.6 6 0.7) (0.7 6 0.2) (3.20) þ11%

Hippocampus 4.7 6 1.1 1.8 6 0.5 0.7 6 0.1 3.31(5.2 6 1.3) (1.8 6 0.5) (0.5 6 0.1) (3.39) þ2%

Occipital with Temporal cortex 6.7 6 2.6 2.9 6 1.2 0.6 6 0.1 2.91(3.0 6 0.8) (1.1 6 0.3) (0.5 6 0.1) (3.23) þ10%

Inferior colliculi 41.3 6 8.4 3.0 6 0.6 0.6 6 0.4 14.4(38.8 6 9.7) (3.0 6 0.8) (0.8 6 0.6) (13.7) �5%

Superior colliculi 29.0 6 7.9 3.1 6 0.9 0.5 6 0.4 9.85(22.8 6 9.6) (2.5 6 1.1) (0.7 6 0.6) (9.82) �0%

Pons with medulla 7.8 6 2.0 2.1 6 0.5 0.6 6 0.1 4.31(7.5 6 1.8) (1.9 6 0.5) (0.5 6 0.2) (4.45) þ3%

Cerebellum 2.4 6 0.8 2.6 6 0.9 0.5 6 0.1 1.42(2.0 6 0.9) (2.1 6 1.0) (0.4 6 0.1) (1.35) �5%

Saturation parameters (with SE estimated from the fits) for control rats and rats pretreated with GBL, the latter shown in parentheses. Rats were given[11C]FLB 457 with or without an additional dose of FLB 457 and tissue:plasma radioactivity concentration ratios were measured at 60 min. Units for Bmax andED50 are nmol/kg. Maximal binding ratio was estimated as (Bmax/ED50) þ NS. Percent occupancy due to dopamine was calculated as 100*(ratio GBL � ratio con-trol)/ratio GBL.

Fig. 5. Maximal regional tissue:plasma radioactivity concentra-tion ratios estimated at zero occupancy by cold FLB 457, from doseresponse curves such as those shown in Figure 4 (and listed inTable II), compared with BP values estimated from kinetic datasuch as those shown in Figure 1 (and listed in Table I). The annota-tion is as for Table I. The linear correlation coefficient (R), calcu-lated by least squares was 0.948 (P < 0.001) and the fitted line isdefined by (y ¼ 1.23x þ 1.09).

167EXTRASTRIATAL BINDING OF [11C]FLB 457

reduction in Bmax was measured in the neocorticalregions (mean 6 SD ¼ 3.6 6 0.6 nmol/kg comparedwith 7.5 6 0.8 nmol/kg). Considering only these tis-sues, two-way ANOVA showed a significant effect ofdrug (P ¼ 0.004), with each region significantlyreduced compared with control (P < 0.05, with Bonfer-roni correction for multiple comparisons). A statisti-cally significant reduction in ED50 was obtained usingtwo-way ANOVA and all regions, but the statistical sig-nificance again increased, from P ¼ 0.015 to P ¼ 0.009,when only the neocortical tissues were considered. Forthe latter, the mean 6 SD ED50 values (n ¼ 3) in con-trol and GBL-treated rats were 3.1 6 0.5 nmol/kg and1.3 6 0.3 nmol/kg, respectively, although, in thisinstance, no individual region showed a significantreduction after Bonferroni correction. The estimate formaximal tissue:plasma ratio increased slightly, but sig-nificantly, from 3.0 6 0.2 to 3.3 6 0.2 (2-way ANOVA,P ¼ 0.001) in cortical regions, where percent occupancyof the D2-like receptors by endogenous DA was calcu-lated to be �10% (100*(GBL ratio – control ratio)/GBLratio).

Regional values for ED50 and maximal ratio areadditionally presented in Figure 7. Regions showinginsignificant or minimal effects were hypothalamus(Figure 6a), hippocampus, brain stem, colliculi, andcerebellum, the latter illustrated in Figure 6c. The sim-ilarity in the saturation curves lends credence to themethodology, despite relatively high errors in theparameter estimates. In addition to changes in corticalregions, ED50 values were reduced in thalamus andolfactory tubercles, although these were not statisti-cally significant by two-tailed t test.

DISCUSSION

The rationale for the development of a PET or SPETradioligand of sufficiently high affinity and selectivityto enable the quantification of extrastriatal DA D2-likereceptors is given elsewhere (e.g., de Paulis, 2003;Kessler et al., 1991; Soares and Innis, 1999). Although[11C]FLB 457 has been used successfully in clinicalinvestigations of schizophrenia (e.g., Suhara et al.,2002), personality traits (e.g., Yasuno et al., 2001), andpain (Hagelberg et al., 2002), the observed changes inbinding are relatively small and are not always consis-tent. We, therefore, considered it worthwhile to returnto experimental studies in rat to help interpret humanPET data concurrently acquired in our Unit.

The in vivo distribution of [11C]FLB 457 binding inrat brain (BP estimated kinetically over 120 min andshown in Table I) was consistent with the distributionof D2-like receptors previously defined by in vitro auto-radiography, using the agonist ligand, [3H]quinpirole(D3:D2 receptor selectivity �6-fold) (Levant et al., 1992)and the high affinity antagonist ligand, [18F]fallypride(D2:D3 receptor selectivity �6-fold) (Mukherjee et al.,

Fig. 6. Tissue concentration of [11C]FLB 457 in GBL pretreatedrats, 60 min after radioligand injection, as a function of FLB 457dose, for (a) hypothalamus (b) prefrontal cortex and (c) cerebellum.Each datum point is from an individual rat, normalized to plasmaradioactivity. The solid lines represent the best fits to a single-sitebinding model. The dotted lines are the equivalent fits to the controlrats and are from Figure 4.

168 R. AHMAD ET AL.

1999). At 60 min after radioligand injection, theregional uptake in extrastriatal tissues correlatedclosely with BP (Fig. 2), although binding in striatumwas grossly underestimated. This latter finding is con-sistent with the time-to-equilibrium issues already dis-cussed for radiolabeled FLB 457 (Loc’h et al., 1996; Ols-son and Farde, 2001) and for other PET or SPETligands with high affinity (de Paulis, 2003). Since it

was not feasible to obtain radioactivity:time data foreach dose of FLB 457 needed to define saturationcurves, the single 60 min assay time was used for theremainder of the study, with quantification limited toextrastriatal tissues.

Using doses in the range 0.04–2.70 lmol/kg, satura-tion kinetics for [11C]FLB 457 binding in control ratswere adequately described by a single binding sitemodel (Fig. 4). The nonsaturable component of bindingwas similar across all regions and the maximal tissue:plasma ratios estimated from the saturation curvescorrelated closely with the BP distribution measuredkinetically (Fig. 5). Extrastriatal ED50 was estimatedto be in the range 2–3 nmol/kg (Table II). Using a meanvalue of 2.7 nmol/kg, occupancy by stable FLB 457 atthe doses of the PET ligand used in the first control ser-ies of experiments was retrospectively calculated as�20%. The data clearly illustrate the risk of underesti-mating the specific signal when rodents (small bodymass) are used to evaluate putative, high affinity PETtracers, as discussed in Hume et al. (1998). Of interest,the ED50 estimate for the olfactory tubercle was anorder of magnitude higher than in other tissuessampled and explained the relatively low BP comparedwith Bmax estimate for this region.

Using an ED50 of 2.7 nmol/kg and assuming a viableweight extrapolation from rat (�300 g) to man (�70 kg),a human PET dose of �200 MBq at a specific activity of�100 MBq/nmol would result in an extrastriatal recep-tor occupancy of �1%. Thus, tracer kinetics would beachievable in human PET for specific activities of �100MBq/nmol, the latter attainable with current radio-chemistry methodology (Sandell et al., 2000). Althoughthe in vivo estimates of striatal ED50 for [

11C]racloprideappeared approximately equivalent in mouse, rat andman (as discussed in Hume et al. (1998)), this may notbe the case for FLB 457, with differences reflecting, forexample, the free fraction. Sudo et al. (2001) have,however, recommended a human FLB 457 dose of <1.6nmol/body to limit mass effects to below 5%. For sub-jects with normal body weight, they estimate a specificactivity requirement of >116 MBq/nmol for an injectateof 185 MBq (5 mCi), a relationship not dissimilar fromthat predicted from the rat data, despite the referencetissue modeling approach used in the PET analysis. Ifthe slight regional differences in ED50 are real (TableII and Fig. 7), then tracer kinetics may be more diffi-cult to achieve in some extrastriatal tissues (e.g., hypo-thalamus and hippocampus) than in others.

While it is accepted that D2-like receptors arepresent at only a very low density in cerebellum(Camps et al., 1990), the data presented in Figure 4chighlight the fact that [11C]FLB 457 binds to rat cere-bellum in a saturable manner, and that Bmax and ED50

values are comparable with those measured in otherlow binding extrastriatal regions (Table II). The resultsare consistent with other in vivo studies in rat, using

Fig. 7. Effect of GBL pretreatment on (a) ED50 and (b) maximaltissue:plasma ratio for each of the extrastriatal regions dissected.The annotation is as listed in Table I and the data are detailedTable II. The ED50 values for the cortical regions (prefrontal cortex,frontal with parietal cortex, occipital with temporal cortex) are clus-tered below the line of identity (shown as the solid line for eachparameter).

169EXTRASTRIATAL BINDING OF [11C]FLB 457

[18F]fallypride (Mukherjee et al., 1999), and in vitrowork using [125I]epidepride, the iodo-analog of FLB457 (Kessler et al., 1991) and suggest caution in usingcerebellum in a reference tissue modeling approach toquantification of kinetic data. Here, we report a 2- to 3-fold difference in the ‘nonspecific’ component of re-gional binding estimated from saturation kinetics andcerebellum binding measured at ‘‘tracer’’ doses ofligand (Table II), predicting underestimations of extra-striatal BP and occupancy using cerebellum as a refer-ence region. Normalizing to the individual cerebellumradioactivity concentration gave ED50 estimates thatwere several fold higher than those listed in Table II(data not shown). Since FLB 457 has equally highaffinities for the D2 and D3 receptor subtypes (Halldinet al., 1995), specific binding to the D3 receptor subtypein cerebellum cannot be excluded. While the primelocation of the D3 receptor is in ‘‘limbic’’ regions, with apostulated role in DA-influenced complex behaviors(Sokoloff and Schwartz, 1995), subtype selective ligandbinding (e.g., Bancroft et al., 1998; Levesque et al.,1992) has been reported in lobules 9 and 10 of the ratcerebellum, with a possible functional role in the regu-lation of locomotor activity (Barik and Beaurepaire,1996).

The use of PET or SPET radioligand binding to D2-like receptors to measure fluctuations in synaptic DAin ‘‘ligand activation’’ studies is reviewed in Laruelle(2000). As briefly described earlier, the in vivo bindingof a radioligand can be moderated by competition withthe neurotransmitter, according to the following rela-tionship:

BP ¼ Bmax=ðligandKd�ð1þ ½DA�=KiÞÞ

as detailed in Ginovart et al. (1997).While it is accepted that the ability to monitor extra-

striatal DA would be of great interest in clinical PET,preclinical investigations as to the susceptibility of[11C]FLB 457 binding to an increase in DA have notbeen entirely consistent (Chou et al., 2000; Okauchiet al., 2001). Recently, however, Aalto et al. (2005) havereported significant reductions (�10%) in [11C]FLB 457binding in selected cortical regions during workingmemory and sustained attention tasks in man. Ofinterest, DA has a high affinity at D3 receptors (Sokol-off et al., 1992; Sokoloff and Schwartz, 1995), related toa minimal difference in affinities of the G proteincoupled and uncoupled states of the D3 receptor foragonists (Burris et al., 1995; Levant, 1997). Based onin vitro and ex vivo autoradiographic findings, Schotteet al. (1992, 1996) have provided evidence for tightbinding of endogenous DA to the D3-subtype, with invivo occlusion of the receptor by neurotransmitter, to amuch greater extent than at the D2-subtype. Thus,prior in vivo depletion of monoamines or preincubationof rat brain sections was needed to obtain quantitative

radioligand binding in the islands of Calleja andnucleus accumbens, although binding in cerebellarlobules 9/10, an area lacking endogenous DA, was notaffected.

In the current study, we observed only a minimaleffect of GBL pretreatment on both maximal bindingand ED50 values for extrastriatal binding of [11C]FLB457 (Table II). Analysis of variance did, however, showa small but statistically significant reduction in ED50

in the neocortical region group (n ¼ 3). Paradoxically,the cortical Bmax was also decreased in the DA-depletedrats. We have no full explanation for this finding.While it could be related to a DA-dependent conforma-tional change in the receptor (Bacopoulus, 1981), thepossibility of covariation of variables in the mathemati-cal model cannot be excluded. Despite the apparentreduction in Bmax, the tracer in vivo signal (a measureof receptor number/apparent affinity), measured as themaximal tissue:plasma ratio, was significantly in-creased (P < 0.02, by analysis of variance) and percentoccupancy of the D2-like receptors by endogenous DAwas calculated to be �10% (100*(GBL ratio – controlratio)/GBL ratio).

Previously, Delforge et al. (2001) have reported noeffect of reserpine on extrastriatal binding of[76Br]FLB 457 in baboons. In human SPET, however,depletion of DA by pretreatment with a-methyl-para-tyrosine, resulted in a slight increase in the BP of[123I]epidepride in temporal cortex (Fujita et al. 2000)and the authors concluded higher receptor occupancyin cortex compared with thalamus. In general terms,the results from our study are consistent with a low,but measurable baseline occupancy in neocortical tis-sues. The low basic tonic release compared with thatexpected in striatum is consistent with a lower numberof DA terminals and lower synaptic concentrations ofDA. Using microdialysis to monitor extracellular levelsof DA, Moghaddam and Bunney (1990), for example,have measured a baseline value of 0.28 fmol/lml in ratmedial prefrontal cortex compared with 3.3 fmol/lml instriatum.

We were unable to detect a statistically significantreduction in ED50 in cerebellum, despite the postulatedpresence of the D3-subtype (high DA affinity). Again,this is consistent with lack of endogenous DA (Schotteet al., 1996) and/or the possible extrasynaptic localiza-tion of the D3 receptor in cerebellum (Sokoloff andSchwartz, 1995). As ligands with a higher affinity dif-ferential (D3 > D2) become available (e.g., Sovago et al.,2004), then the contribution of D3-subtype binding tothe in vivo regional binding of nonsubtype selectivePET ligands such as [11C]FLB 457 could be easilydetermined experimentally, using methodologies suchas those described above.

In summary, the regional binding data obtained inrats confirmed the need for high specific activity prepa-rations of [11C]FLB 457 in order to satisfy tracer

170 R. AHMAD ET AL.

kinetics, and highlighted a further potential quantita-tive inaccuracy if tracer binding in cerebellum isassumed to represent nonspecific distribution. Consis-tent with earlier preclinical studies using nonhumanprimates, the current work demonstrates only a lowextrastriatal basal occupancy by DA, but does suggesta finite possibility of using PET with [11C]FLB 457 toindex DA fluctuations in cortical regions.

ACKNOWLEDGMENTS

The authors thank members of the radiochemistryproduction team at Hammersmith Imanet Ltd. forfacilitating these studies, and Dr. Marie-Claude Asse-lin for many searching discussions.

REFERENCES

Aalto S, Bruck A, Laine M, Nagren K, Rinne JO. 2005. Frontal andtemporal dopamine release during working memory and attentiontasks in healthy humans: a positron emission tomography studyusing the high-affinity dopamine D2 receptor ligand [11C]FLB457. J Neurosci 25:2471–2477.

Bacopoulos N. 1981. Acute changes in the state of dopamine receptors;in vitro monitoring with 3H-dopamine. Life Sci 29:2407–2414.

Bancroft GN, Morgan KA, Flietstra RJ, Levant B. 1998. Binding of[3H]PD 128907, a putatively selective ligand for the D3 dopaminereceptor, in rat brain: a receptor binding and quantitative autora-diographic study. Neuropharmacol 18:305–316.

Barik S, de Beaurepaire R. 1996. Evidence for a functional role ofthe dopamine D3 receptors in the cerebellum. Brain Res 737:347–350.

Burris KD, Pacheco M, Filtz TM, Kung ,M-P, Kung HF, Molinoff PB.1995. Lack of discrimination by agonists for D2 and D3 dopaminereceptors. Neuropsychopharmacol 12:335–345.

Camps M, Kelly PH, Palacios JM. 1990. Autoradiographic localiza-tion of dopamine D1 and D2 receptors in the brain of severalmammalian species. J Neural Transm 80:105–127.

Chou YH, Halldin C, Farde L. 2000. Effect of amphetamine onextrastriatal D2 dopamine receptor binding in the primate brain:a PET study. Synapse 38:138–143.

de Paulis T. 2003. The discovery of epidepride and its analogs ashigh-affinity radioligands for imaging extrastriatal dopamine D2

receptors in human brain. Curr Pharm Des 9:673–696.Delforge J, Bottlaender M, Pappata S, Loc’h C, Syrota A. 2001.

Absolute quantification by positron emission tomography of theendogenous ligand. J Cereb Blood Flow Metab 21:631–630.

Farde L, Hall H, Ehrin E, Sedvall G. 1986. Quantitative analysis ofD2 dopamine receptor binding in the living human brain by PET.Science 231:258–261.

Farde L, Eriksson L, Blomquist G, Halldin C. 1989. Kinetic analysis ofcentral [11C]raclopride binding to D2-dopamine receptors studied byPET—a comparison to the equilibrium analysis. J Cereb Blood FlowMetab 9:696–708.

Fujita M, Verhoeff NPLG, Varrone A, Zoghbi SS, Baldwin RM, Jat-low PA, Anderson GM, Seibyl JP, Innis RB. 2000. Imaging extra-striatal dopamine D2 receptor occupancy by endogenous dopaminein healthy humans. Eur J Pharmacol 387:179–188.

Ginovart N, Farde L, Halldin C, Swahn C-G. 1997. Effect of reser-pine-induced depletion of synaptic dopamine on [11C]raclopridebinding to D2-dopamine receptors in the monkey brain. Synapse.25:321–325.

Hagelberg N, Martikainen IK, Mansikka H, Hinkka S, Nagren K,Hietala J, Scheinnin H, Perrovaara A. 2002. Dopamine D2 recep-tor binding in the human brain is associated with the response topainful stimulation and pain modulatory capacity. Pain 99:273–279.

Halldin C, Farde L, Hogberg T, Mohell N, Hall H, Suhara T, Karls-son P, Nakashima Y, Swahn CG. 1995. Carbon-11-FLB 457: aradioligand for extrastriatal D2 Dopamine receptors. J Nucl Med36:1275–1281.

Hogberg T. 1991. Novel substituted salicylamides and benzamidesas selective dopamine D2-receptor antagonists. Drugs Future16:333–357.

Hume SP, Myers R, Bloomfield P, Opacka-Juffry J, Cremer J, AhierRG, Luthra SK, Brooks DJ, Lammertsma AA. 1992. Quantitationof carbon-11-labeled raclopride in rat striatum using positronemission tomography. Synapse 12:47–54.

Hume SP, Brown DJ, Ashworth S, Hirani E, Luthra SJ, Lam-mertsma AA. 1997. In vivo saturation kinetics of two dopaminetransporter probes measured using a small animal positron emis-sion tomography scanner. J Neurosci Meth 76:45–51.

Hume SP, Gunn RN, Jones T. 1998. Pharmacological constraintsassociated with positron emission tomographic scanning in smalllaboratory animals. Eur J Nucl Med 25:173–176.

Kessler RM, Ansari MS, Schmidt DE, Paulis DT, Clanton JC,Innis R, al-Tikriti M, Manning RG, Gillespie D. 1991. High affin-ity dopamine D2 receptor radioligands. 2. [125I] epidepride, apotent and specific radioligand for the characterization of striataland extrastriatal dopamine D2 receptors. Life Sci 49:617–628.

Kessler RM, Whetsell WO, Ansari MS, Votow JR, de Paulis T, Clan-ton JA, Schmidt DE, Mason NS, Manning RG. 1993. Identificationof extrastriatal dopamine D2 receptors in post mortem humanbrain with [125I]epidepride. Brain Res 609:237–243.

Kohler C, Hall H, Ogren S-O, Gawell L. 1985. Specific in vitro andin vivo binding of 3H-raclopride, a potent substituted benzamidedrug with high affinity for dopamine D-2 receptors in the ratbrain. Biochem Pharmacol 34:2251–2259.

Lammertsma AA, Bench CJ, Hume SP, Osman S, Gunn K, BrooksDJ, Frackowiak RSJ. 1996. Comparison of methods for analysis ofclinical [11C]raclopride studies. J Cereb Blood Flow Metab 16:42–52.

Laurelle M. 2000. Imaging synaptic neurotransmission with in vivobinding competition techniques: A critical review. J Cereb BloodFlow Metab 20:423–451.

Levant B. 1997. The dopamine D3 receptor: neurobiology and poten-tial clinical relevance. Pharmacol Rev 49:231–252.

Levant B, Grigoriadis DE, DeSouza EB. 1992. Characterization of[3H]quinpirole binding to D2-like dopamine receptors in rat brain.J Pharmacol Exp Ther 262:929–935.

Levesque D, Diaz J, Pilon C, Martres M-P, Giros B, Souil E, Schott D,Morgat JL, Schwartz JC, Sokoloff P. 1992. Identification, character-ization, and localization of the dopamine D3 receptor in rat brainusing 7-[3H]hydroxy-N, N-di-n-propyl-2-aminotetralin. Proc NatlAcad Sci USA 89:8155–8159.

Lo’ch C, Halldin H, Bottlaender M, Swahn C-G, Moresco R-M,Maziere M, Farde L, Maziere B. 1996. Preparation of [76Br]FLB457 and [76Br]FLB 463 for examination of striatal and extrastria-tal Dopamine D-2 receptors with PET. Nucl Med Biol 23:813–819.

Maitre M. 1997. The g-hydroxybutyrate signalling system in brain:organization and functional Implications. Prog Neurobiol 51:337–361.

Moghaddam B, Bunney BS. 1990. Acute effects of typical and atypi-cal antipsychotic drugs on the release of dopamine from prefrontalcortex, nucleus accumbens, and striatum of the rat: an in vivomicrodialysis study. J Neurochem 54:1755–1760.

Mukherjee J, Yang Z-Y, Brown T, Lew R, Wernick M, Ouyang X,Yasillo N, Chen CT, Mintzer R, Cooper M. 1999. Preliminaryassessment of extrastriatal dopamine D-2 receptor binding in therodent and nonhuman primate brains using the high affinityradioligand, 18F-fallypride. Nucl Med Biol 26:519–527.

Okauchi T, Suhara T, Maeda J, Kawabe K, Obayashi S, Suzuki K.2001. Effect of endogenous dopamine on extrastriatal [11C]FLB-457 binding measured by PET. Synapse 41:87–95.

Olsson H, Farde L. 2001. Potentials and pitfalls using high affinityradioligands in PET and SPET determinations on regional druginduced D2 receptor occupancy—a simulation study based onexperimental data. NeuroImage 14:936–945.

Sandell J, Langer O, Larsen P, Dolle F, Vaufrey F, Demphel S, Crou-zel C, Halldin C. 2000. Improved specific radioactivity of the PETradioligand [11C]FLB 457 by the use of the GE Medical SystemsPETtrace Mel MicroLab. J Labelled Compd Radiopharm 43:331–338.

Schotte A, Janssen PFM, Gommeren W, Luyten WHLM, Leysen JE.1992. Autoradiographic evidence for the occlusion of rat braindopamine D3 receptors in vivo. Eur J Pharmacol 218:373–375.

Schotte A, Janssen PF, Bonaventure P, Leysen JE. 1996. Endoge-nous dopamine limits the binding of antipsychotic drugs to D3

receptors in the rat brain: a quantitative autoradiography study.Histochem J 28:791–799.

Soares JC, Innis RB. 1999. Neurochemical brain imaging investiga-tions of schizophrenia. Biol Psychiatry 46:600–615.

Sokoloff P, Schwartz J-C. 1995. Novel dopamine receptors half adecade later. Trends Pharmacol Sci 16:270–275.

Sokoloff P, Andrieux M, Besacon R, Pilon C, Martres MP, Giros B,Schwartz J-C. 1992. Pharmacology of human dopamine D3 recep-

171EXTRASTRIATAL BINDING OF [11C]FLB 457

tor expressed in a mammalian cell line: comparison with D2

receptor. Eur J Pharmacol 225:331–337.Sovago J, Farde L, Halldin C, Langer O, Laszlovszky I, Kiss B,

Gulyas B. 2004. Positron emission tomographic evaluation of theputative dopamine-D3 receptor ligand, [11C]RGH-1756 in the mon-key brain. Neurochem Int 45:609–617.

Sudo Y, Suhara T, Inoue M, Ito H, Suzuki K, Saijo T, Halldin C,Farde L. 2001. Reproducibility of [11C]FLB 457 binding in extra-striatal regions. Nucl Med Commun 22:1215–1221.

Suhara T, Sudo Y, Okauchi T, Maeda J, Kawabe K, Suzuki K,Okubo Y, Nakashima Y, Ito H, Tanada S, Halldin C, Farde L. 1999.Extrastriatal dopamine D2 receptor density and affinity in thehuman brain measured by 3D PET. Int J Neuropsychopharmacol2:73–82.

Suhara T, Okubo Y, Yasuno F, Sudo Y, Inoue M, Ichimiya T, Naka-shima Y, Nakayama K, Tanada S, Suzuki K, Halldin C, Farde L.2002. Decreased dopamine D2 receptor binding in the anterior cin-gulate cortex in schizophrenia. Arch Gen Psychiatry 59:25–30.

Talvik M, Nordstrom A-L, Olsson H, Halldin C, Farde L. 2003.Decreased thalamic D2/D3 receptor binding in drug-naıve patientswith schizophrenia: a PET study with [11C]FLB 457. Int J Neuro-psychopharmacol 6:361–370.

Vessotskie JM, Kung M-P, Chumpradit S, Kung HF. 1997. Quantita-tive autoradioraphic studies of dopamine D3 receptors in rat cere-bellum using [125I]S(-)5-OH-PIPAT. Brain Res 778:89–98.

Wolf ME, Roth RH. 1987. Dopamine autoreceptors. In: Crees I,Fraser CM, Alan R, editors. Dopamine Receptors. New York: Liss.p 45–96.

Wolfson LI, Sakurada O, Sokoloff L. 1977. Effects of g-butyrolactoneon the local cerebral glucose utilization in the rat. J Neurochem29:777–783.

Yasuno F, Suhara T, Sudo Y, Yamamoto M, Inoue M, Okubo Y,Suzuki K. 2001. Relation among dopamine D2 receptor binding,obesity and personality in normal human subjects. Neurosci Lett300:59–61.

172 R. AHMAD ET AL.