(6,7). Recently,theuseof PEThasallowedin vivoimagingof the regional brain distribution of 5-HT2A receptors (8).However, PET radioligands for 5-HT2A receptor imaging,such as “C-3-N-methyl-spiperone(NMSP) (9) and [18F]-setoperone (10), lack specificity, binding with moderate tohigh affinity to dopamine D2 receptors. As a result, considerable effort has been directed to the development of moreselective 5HT2A ligands for PET (11,12).

[‘8F}altanserin(11, 13), a 4-fluorobenzoyl-piperidine derivative structurally related to ketanserin and setoperone, hasemerged as a promising PET ligand for 5-HT2A receptors.Compared with previous 5-HT2A radiotracers, [‘8F]altanserin yields higher ratios of specific-to-nonspecific bindingand greater selectivity for 5-HT2A receptors compared withother monoamine receptors (14). Altanserin has more than a

20-fold greater selectivity for 5-HT2A than any other 5-HTreceptor subtype, including the 5-HT2C, which has significant homologies to 5-HT2A (15). Kinetic modeling with[‘8F]altanserinhas previously been used in initial humanstudies (16—19),but may be hampered by the presence ofnonspecifically bound lipophilic radiometabolites observedin plasma after intravenous administration (15,20,21), resulting in the need to identify a large number of rate constants.

The purpose of this study was to test the feasibility of aconstant infusion paradigm for equilibrium modeling of[‘8F]altanserinwith PET. Because of problems associatedwith kinetic analysis of single bolus experiments, weimplemented an alternative paradigm based on constantinfusion of the radiotracer. This paradigm creates andmaintains a prolonged state of equilibrium at the level of thereceptors, allowing direct measurement of the distributionvolumes of the tracer (22,23). The presence of lipophilicradiometabolites poses a challenge for an equilibrium approach as well, but the modeling is considerably simpler.That is, if the metabolites are inactive (15,21) and distributeuniformly in brain, then subtraction of background from atarget region at the time of equilibrium provides specificreceptor binding, corrected for the presence of metabolites.Other advantages of an equilibrium approach include theabsence of arterial plasma sampling and the relatively short

[18F]altansennhas been usedto labelserotonin5-HT@receptors, whichare believedto be importantin the pathophysiologyofschizophreniaanddepression.The purposeof thisstudywastotestthefeasibilityofa constantinfusionparadigmforequilibriummodeling of [18F]altanserin with PET. Kinetic modeling with[18F]altansennmay be hamperedby the presenceof lipophilicradiometabolitesobservedin plasmaafterintravenousadministration. Methods: Eight healthy volunteers were injected with[18F]altanserinas a bolus(208 ±9 MBq[5.62 ±0.25 mCi])plusconstantinfusion(65 ±3 MBq/h[1.76 ±0.08 mCi/h])rangingfrom 555 to 626 mm (615 ±24 mm) after injection. PETacquisitions (10—20mm) and venous blood sampling wereperformedevery 30—60mm throughoutthe infusionperiod.Results:Linearregressionanalysisrevealedthattime—activitycurves for both brain activity and plasma [18l9altansennandmetaboliteconcentrationsstabilizedafter about 6 h. This permitted equilibriummodelingand estimationof V3' (ratioof specificuptake[cortical—cerebellar]to totalplasmaparentconcentrationafter 6 h). Valuesof V3' rangedfrom 1.57 ±0.38 for anteriorcingulate cortex to 1.02 ±0.39 for frontal cortex. The bindingpotentialV3 (ratio of specificuptake to free plasma parentconcentrationafter6 h,usinggroupmeanf1)wasalsocalculatedandrangedfrom169 ±41 foranteriorcingulatecortexto 110 ±42 forfrontalcortex. From6 h onward, the rate of change for V3'and V3was only1.11 ±I .69 %Th.Conclusion:These resultsdemonstratethefeasibilityofequilibriumimagingwith[18F]altansenn over more than 5 radioactivehalf-livesand suggestamethodto overcomedifficultiesassociatedwith lipophilicradiolabeledmetaboiites.The stabilityinV3andV3'onceequilibriumisachievedsuggeststhata singlePET acquisitionobtainedat 6 hmayprovidea reasonablemeasureof5-HT@receptordensity.KeyWords:[18F]altanserin;5-HT@receptor;serotonin;PETJ NucIMed2000;41:234—241

n vitro studies on human brain autopsy material haverevealed alterations in serotonin 5-HT2A receptors in avariety of conditions, including depression/suicide (1,2),schizophrenia (3,4), aging (5), and Alzheimer's disease

Received Jul. 27, 1998; revision accepted Jul. 30, 1999.Forcorrespondenceor reprintscontact:christopherH.vanDyck,MD,Yale

University SchoolotMedicine, CB 2041, 333 Cedar St., New Haven, CT 06510.

234 THE JOURNALOF NUCLEAR MEDICIr@ir@•Vol. 41 •No. 2 •February 2000

PET Quantification of 5-HT2AReceptors in theHuman Brain: A Constant Infusion Paradigm with[1 8F]Altanserin

Christopher H. van Dyck, Ping-Zhong Tan, Ronald M. Baldwin, Louis A. Amici, Pradeep K. Garg, Chin K. Ng,Robert Soufer, Dennis S. Charney, and Robert B. Innis

Departments ofPsychiatry and Diagnostic Radiology, Yale University School ofMedicine, New Haven; and West Haven Veterans'Affairs Medical Center@West Haven, Connecticut

by on June 30, 2020. For personal use only. jnm.snmjournals.org Downloaded from

imaging time (20—40mm), once the methodology has beenvalidated for clinical studies.

MATERIALSAND METHODS

RadiopharmaceuticalPreparationandPurity[‘tF]altanserinwas synthesized in a complete remote control

system (24) modified from a method previously described (11).Theprecursor nitro-altanserin, 3-(2-(4-(4-nitrobenzoyl)-l-piperidinyl)-ethyl)-l,2-dihydro-2-thioxo-4-quinazolinone, was prepared fromcommercial reagents with >99.0% purity (25). For these 8 PET

studies, 2057 ±322 MBq (55.6 ±8.7 mCi; with these andsubsequent data expressed as mean ±SD, unless otherwisespecified; SEM is used in Figs. 1, 2, and 4 for legibility)[‘8Flaltanserinwere produced, with an overall radiochemical yieldof 10.3% ±2.5% at the end ofsynthesis. Radiochemical purity was99.0% ±I.3% and specific activity 46,546 ±18,981 MBq/pmol(I 258 ±513 mCi4imol) (range, 22,533—75,591MBq/pmol [609—2043 mCi/pmol]) at the time of injection and start of infusion, asdetermined by high-performance liquid chromatography (HPLC)(EM RPSelect-B [EMScience, Gibbstown,NJ];250 X 4.6 mm, 10@.im,tetrahydrofuran-to-sodiumacetatebuffer,35:65 [pH5.0] at 2.5mL/min). The synthesis time was 112 ±9 mm (range, 97—124mm). The total mass dose of [‘8F]altanserinadministeredduringthe bolus plus constant infusion study was 0.14 ±0.05 pg/kg(range, 0.10—0.24pg/kg). Sterility was confirmed by lack of growthin 2 media: fluid thioglycollate and soybean-casein digest (26).Apyrogenicity was confirmed by the LAL test (Endosafe, Charleston, SC).

StudyPopulationThe study population included 8 healthy volunteers (3 men, 5

women; age range, 25—42y; mean age, 32.1 ±5.6 y) and ranged inweight from 43 to 82 kg (63.8 ±12.5 kg). Volunteers underwentclinical examinations by a research psychiatrist to exclude anyneurological or psychiatric disease or alcohol or substance abuse.Screening procedures included a physical and neurological examination, electrocardiography, serum chemistries, thyroid functionstudies, complete blood cell count, urinalysis, and urine toxicologyscreen.In addition,eachvolunteerunderwentbrain MRI that wasread as normalby a neuroradiologist.No volunteer,according toself-report, had taken centrally active medication within I mo ofPETscanning.All volunteersgave writteninformedconsent to theresearch protocol approved by the local human investigationcommittee.

PETDataAcquisitionBased on preliminary studies in 5 healthy volunteers (data not

shown), a bolus/infusion (hourly infusion rate) ratio of 3.2 wasselected for the present analysis. [‘8F]altansenn(208 ±9 MBq[5.62 ± 0.25 mCi]) was injected intravenously (through theconstant infusion line) as a single bolus over 30 s at approximately12 noon, and a constant infusion of 65 ±3 MBq/h (1.76 ±0.08mCi/h)was startedandranin durationfrom555 to 626 mm (615 ±24 mm). Thus, the total administered dose (incorporating decay ofthe isotope) was 377 ±19 MBq (10.2 ±0.5 mCi). The actualbolus/infusion ratio for the 8 volunteers was 3.20 ±0.02.

Volunteers were placed in the Posicam 6.5 camera (PositronCorp., Houston, TX) at the Yale/Veterans'Affairs PET Center, withhead positioned by laser along the canthomeatal line and immobilized withstraps.Repositioningafterperiodsoutofthe scannerwasaccomplished by realigning the laser lights with marks placed

along the canthomeatal line and glabella. The Posicam 6.5 is a21-slice camera with 5.125-mm interslice distance. The resolutionof the camera is 5.8 mm in-plane and 11.9 mm in the z axis.In-planeresolutionafter filteringwith the Butterworthfilter is 7.5mm. The sensitivity of the camera measuredin a 20-cm-diametercylinder phantom is 165 kcts/s/pCi/mL (27). Five fiducial markersfilled with 0.5—1.0pCi ‘8Fwere attached to the skin along thecanthomeatalplane to identify this plane during image analysis.PETwas performedthroughoutthe [‘8F]altanserininfusion periodaccordingto the following protocol: 10-mm frame mode acquisitions starting at approximately 45, 90, 120, 180, 240, and 300 mmafter injection; then 20-mm frame mode acquisitions starting atapproximately360, 390, 450, 480, 510, 540, 570, and600 mm afterinjection. PET and plasma data were more densely sampled from6 h onward to evaluate the assumptions of the equilibrium modelmore reliably late in the study.To minimize repositioningerrors,volunteersremainedin the scannerfor up to 3 h at a time, taking2breaks out of the scanner during the experiment.

PETimageAnalysisPETimages werereconstructedusinga deconvolutiontechnique

for scatter correction and a Butterworthfilter (cutoff, 0.09/mm;power factor, 10) anddisplayedas a 256 X 256 matrix(pixel size,1.7 X 1.7 mm; slice thickness, 5.125 mm; voxel volume, 14.81mm3). Attenuation correction was performed assuming uniformattenuation equal to that of water (attenuation coefficient, p =0.096/cm) within an ellipse drawn around the skull as identified bythe markers.A cylindrical (20-cm diameter) fluid-filled phantomwas scanned on the day of each study to obtain a calibration factorforeach of the 21 slices forconversionof radioactivityinto unitsofnCi/mL.

Region of interest (ROl) analysis was performed by an investigator using the AnalyzelC image analysis program(CNSoftware,Wilmington, DE) running on a Sun SPARC station 20 (SunMicrosystems, Inc.; Mountain View, CA). ROt analysis wasconducted for 6 cortical structures (frontal, anterior cingulate,orbitofrontal,temporal,parietal,andoccipitalcortices)andcerebellum.These regionswere identifiedby comparisonwith a neuroanatomical atlas (28). ROIs were drawn on the PET image using amouse-driven cursor. The ROl map was drawn individually foreach volunteer with respect to the first PET acquisition and wasapplied uniformly to all subsequent acquisitions for that person.RO!analysiswas performedin 17of the 21 slices, with the inferiorslice chosen to correspond best to the level of the fiducial markers(canthomeatal plane). The superior slice was thus centered approximately8.7 cm above thecanthomeatalplane.

A mean count density (cpm/pixel) was obtained for eachanatomic structure in each acquisition by averaging values acrossall slices in which the structuresappeared.Count densities werethen decay-corrected for the time of injection and expressed innCi/mL to generate brain time—activitycurves.

PlasmaMetaboliteAnalysisAn intravenous line was inserted in the arm opposite the

constant infusion line for obtaining venous blood samples. Venousblood samples were obtained throughout the [‘8F]altanserininfusion period at approximately 30, 60, 120, 180, 300, 360, 390, 420,450, 480, 510, 540, 570, 600, and630 mm afterinjection.

Metabolite analysis was based on a published procedure (16).Plasma(2 mL) was dilutedwith 20 mL 50 mmol/Lacetic acid andtrapped on a C18solid-phase extraction cartridge (Sep-Pak Lite;WatersAssociates, Milford, MA) preconditionedwith methanol

PET QUANTIFICATIONOF 5-HT2A RECEPTORS•van Dyck et al. 235

by on June 30, 2020. For personal use only. jnm.snmjournals.org Downloaded from

andwater,washed with 10 mL 0.1% Et3Nfollowed by 5 mL H20,then eluted with 1 mL MeOH. The methanol eluate was dilutedwithanequal volume H20, andtheentiresolutionwas analyzedbyHPLC on a Waters Novapak C18,3.9 X 300 mm column; mobilephase CH3CN/0.353 mol/L HOAc/0.17 mollL NH.@OAc(pH 4.9)(40:50:10), at a flow rate of 1.0 mLlmin, with in-line ultravioletdetector(291 nm) and flow-throughNa! scintillationdetector/ratemeter.Recoveryofradioactivity using this procedurewas 88.5% ±6.0%, measured by adding tracer to blood and carrying out thesame procedure. For identification of radiometabolites, the radioactive methanol/water solution was mixed with 3 authentic cornpounds (altanserin, altanserinol, and 4-(,p-fluorobenzoyl)piperidine[FBP]) and injected on the HPLC.

Plasma protein binding was measured by ultrafiltrationaspreviously described (29). The binding to plasma proteins wasassumed to be rapid compared with the other processes measured inthese experiments.As a consequence, the free fractionof plasmaparent compound (f1 = free parent/C@,,(t))(30) was assumed to beconstant for all blood samples in a single experiment.

EquilibriumModelEquilibrium analysis was performed using a 3-compartment

model, comprising the plasma compartment (Cr); the intracerebralnondisplaceablecompartment(C2),in which the tracercan be freeor nonspecifically bound;and the specifically boundcompartment(C3).At equilibrium,the venous concentrationof traceris assumedto approximatethe arterialconcentrationandis takenas a measureof C@.The equilibriumdistributionvolume of a compartmenti (V1)is defined as the equilibrium ratio of the tracer concentration in thatcompartment to the free plasma concentration: V =where f1 = free fraction of parent tracer in plasma. V3 is equal tothe bindingpotential(BP), which is theratioofthe receptordensity

@ to thedissociationconstant(KD)(30).The activityconcentrationin a receptor-richcorticalROl at time

t, C@@(t),is the sum of the concentrations of the nondisplaceable[‘8F]altanserin,the specifically bound [‘8F]altanserin,and theradioactive metabolites (which are also assumed to be nondisplaceable):

C@@(t)= C2(t)+ C3(t)+ C,,,..@(t).

Cerebellum has a relatively low concentration of 5-HT2Areceptors (31—33).Ifcerebellum is assumed to be devoid of 5-HT@receptors,then the activity concentrationthereat time t, CCbI(t),isthesumofthe concentrationsofthe nondisplaceable[‘@F}altanserinand the radioactive metabolites:

CCbI(t)= C2(t)+ C,ret(t).

If C2andC@1arefurtherassumedto be uniformthroughoutthebrain,then the specific binding in any receptor-richcorticalregionmay be estimatedas:

C3(t)= C@@(t)—CCb,(t).

The equilibriumvolume of distributionof compartmenti is theratioof the tracerconcentrationin compartmenti to the free tracerconcentration in the plasma at equilibrium, i.e., when no nettransfer between the plasma and compartment i exists:

Vi = C1/FIC@.

Vi, is defined as the ratio of the tracer concentration incompartment i to the total tracer concentration in the plasma at

equilibrium:

vi, = C./C@.

At equilibrium,2 different outcome measures can be derived.V3,theequilibriumvolume of distributionof the specificallyboundcompartment, is an estimate of the BP and is calculated as the ratioof the specific to the plasmafree parentconcentration:

V3 = BP = C3/f1C@.

At tracer radioligand concentration, BP equals Bmss/KD, @flwhich B,,,,,@,is the density of sites and KD is the equilibriumdissociation constant (34). V3' is a similar ratio to the total parent(which therefore does not depend on f1—onlyabout 1% foraltanserin):

V3' = C3/C@.

StatisticalAnalysisThe proposedequilibriummodel relies on several assumptions,

the most importantof which is the stability over time of specifictracer uptake and free plasma parent concentrations. Averagetime—activitycurves for both PET and plasma measurements weregenerated and visually examined to determine when a state ofsustained equilibrium appeared to be achieved. To measure thestability of PET and plasma parameters from this time onward, thehourlyrateofchange foreach parameterwas computedby dividingthe linearregressionslope by the meanvalue of the parameter.

The equilibriummodel was then appliedto the entire sample tocalculate the distribution volumes V3 (C@/f1C@)and V3' (C@/C@).The effect of age on V3andV3'was examinedby Pearson'sproductmoment correlation coefficient (r). Finally, the stability of thedistribution volumes V3 and V3' at equilibrium was assessed bycomputing the hourly rate of change of the distribution volumes foreach volunteer and for the entire sample.

RESULTS

None of the volunteers experienced any subjective orobjective reactions. All vital signs (pulse, blood pressure,and respirations) were recorded before and 15—60mm afterbolus injection and start of constant infusion, and serumchemistry profile, complete blood count, and urinalysis wererepeated the day after the [‘8flaltanserin PET study. Noclinically significant effects on vital signs or laboratorystudies were observed.

Seven of 8 volunteers completed planned [‘8Fjaltanserininfusions of at least 620 mm. Because of a technicalproblem, 1 volunteer completed an infusion of only 555 mm;her data were not used to determine stability of regionalbrain uptake and plasma parent activity but were included inthe final computations of V3 and V31. Inspection of averageplasma metabolite concentrations (Fig. 1A) and averagePET time—activitycurves (Fig. 2) suggested that equilibriumwas attained after approximately 6 h of [18F]altanserininfusion. To examine the stability of plasma and brainparameters after 6 h, the hourly rate of change for eachparameter was computed by dividing the linear regressionslope by the average value of the parameter from 6 to 10.5 hfor the 7 volunteers who completed the full infusionexperiment.

236 ThE JOURNALOFNUCLEARMEDICINE•Vol. 41 •No. 2 •February 2000

by on June 30, 2020. For personal use only. jnm.snmjournals.org Downloaded from

. . .. . *aft@nsenn

.. . 0••- “FBP―

. . A- . altansennoi

A100@

80

60.

40

20-

A

-JE

C.)

C0

L.

CG)UC0C.) 4'@4&:-@

B 70

C@

FiGURE 1. Average plasma metaboliteconcentrations (A) and percentage composition (B) (mean ±SEM used for legibility)after[18F]altanserininfusionin7 volunteersstudiedforat least620 mm.(Onevolunteerwhoseinfusiondurationwas555mmisnotincluded.)Two missingplasma samples(out of 105 total) were handled by usinglinearinterpolationofprecedingandsubsequentsamplesfor that subject.From6 honward, rates of change were low for altanserin (—0.69±5.21 %/@)and major lipophilic metabolite fraction (“FBP―;3.64 ±1.90 %/@),whereas some increase was stillobservedinminorlipophilicmetabolitealtanserinol(5.78 ±9.02 %ft@)Percentageofunchanged[18F]altanserinfell from63% at37 mm after injection and stabilized atapproximately45%from6 honward.

I'

PlasmaMetabolfteConcentrationsAfter[‘8F]AltanserinInfusion

Three major plasma components were observed in theseexperiments. [‘8F]altanserinparent (HPLC retention time tR,10.7 ±0.3 mm) and the ketone reduction product [‘8F]altanserinol(tR,8.1 ±0.2 mm) were identifiedby comparisontoauthentic standards co-injected with the radioactive plasmasamples. The third metabolite fraction had retention time (tR,6.2 ±0.1 mm) close but not identical to that of authenticFBP (tR, 5.6 ±0.1 mm), the product of N-dealkylation onthe piperidine ring. Results from other studies (15,20,21)suggest that this fraction actually consists of a mixture ofmetabolites, some of which may be secondary metabolitesof FBP. In the present analysis, these additional metabolitesare presumed to be contained in the FBP peak on HPLC andare included in the “FBP―time—activitycurves.

Average plasma metabolite concentrations after [‘8F]altanserin infusion in the 7 volunteers studied for at least 620 mm

are displayed in Figure lÀ.The average time—activitycurvefor [18F]altanserinappeared to level offby approximately 6 hafter injection. From 6 h onward, mean rates of change werelow for altanserin (0.69 ±5.21 %/h) and the major lipophilicmetabolite fraction (“FBP―; 3.64 ± 1.90 %/h), whereassome increase was still observed in the minor lipophilic

metabolite altanserinol (5.78 ±9.02 %/h).The kinetics ofunmetabolized [‘8F]altanserinand metabo

lites as a fraction of the plasma activity are shown in FigurelB. The percentage of unchanged [18F]altanserin fell from63% at 37 mm after injection and stabilized at approximately 45% from 6 h onward.

PETTime—ActivityCurvesAfter(@9AItanserIninfusionRepresentative [‘8FjaltanserinPET images from early and

late time points are displayed in Figure 3. [‘8F]altanserin

uptake is high and relatively uniform in all cerebral cortical

PET QUANTIFICATIONOF 5@'T2A RECEPTORS•van Dyck et al. 237

. . .- - *altansenn

.. •0@-- “FBP―J ..A-

w

0 100 200 300 400 500 600Minutes

..@..+..

0 1ÔÔ@2Ô@@@ àóóióó‘êóóMinutes

by on June 30, 2020. For personal use only. jnm.snmjournals.org Downloaded from

regions and low in basal ganglia, thalamus, and cerebellum.

Average PET time—activity curves after [18flaltanserininfusion in the 7 volunteers studied for at least 620 mm aredisplayed in Figure 2. From 6 h onward, rates of changewere low for receptor-rich cortical regions (2.46 ± 1.03%Ih), receptor-poor cerebellum (3.45 ± 2.10 %/h), andspecifically bound (cortex—cerebellum) tracer uptake (1 .16 ±2.00 %/h) (Table 1).

EquilibriumOutcomeMeasuresThe outcome measures derived at equilibrium (averaging

each volunteer's data from 6 to 10.5 h after injection) forcortical brain regions in all 8 volunteersare displayed inTable 2. Ratios of total-to-nonspecific (cerebellar) bindingbetween 6 and 10.5 h after injection were 1.97 ±0.16 for thestructure of highest uptake (anterior cingulate) and 1.75 ±0.15 for wholecortex.Valuesof V3' (C@/C@)rangedfromahigh of 1.57 ±0.38 for anterior cingulate cortex to 1.02 ±0.39 for frontal cortex. Measurements of f1 were used tocalculate V3 (C@/f1C@)as an estimate of the BP. Values of f1ranged from 0.36% to 2.00% (0.93% ±0.63%). V3 wascomputed using both group mean and individual values of f1.

Values of V3' (and therefore V3. using a mean f1) weresignificantly correlated with age in this small sample forwhole cortex (r = —0.73;P < 0.05) as well as orbitofrontalcortex (r = —0.79;P < 0.05), frontal cortex (r = —0.79;P < 0.05), anterior cingulate cortex (r = —0.70;P < 0.05),

FIGURE2. AveragePETtime-activitycurves(mean±SEMusedfor legibility)after [18F]altanserininfusionin 7 volunteersstudiedfor at least 620 mm. (One volunteerwhose infusiondurationwas555 mmisnotincluded.)ThreemissingPETscans(out of 98 total) were handled using linear interpolationofprecedingand subsequentsamplesfor that subject.Cortexrepresentscompositeofall6 corticalregionsanalyzed.From6 honward, rates of change were low for receptor-richcorticalregions(2.46 ±I .03 %/@),receptor-poorcerebellum(3.45 ±2.10 0k/h),and specificallybound(cortex—cerebellum)traceruptake(1.16±2.00 %Ih).

A

FIGURE3. TransaxiaiPET imagesof[18F]altanserinuptake in healthy volunteerat early and late time points:(A) 10-mmacquisitionstarting52 mm after injectionand(B)40-mmacquisitionstarting376 mmafter injection. Uptake of [18F]altanserinisconsistentwithknowndistributionof5-HT@receptorsasseeninautoradiographicstudies,withhighandrelativelyuniformuptakein all cerebralcortical regions,low uptakeinbasal ganglia and thalamus, and muchloweruptakeincerebellum.Slicesare situated approximately80 (high cortex), 45(thalamusand basalganglia),and 10 mm(cerebellum)above canthomeatalplane.Laterimagesare fromsinglelongacquisition obtained at equilibrium, as could beusedinclinicalstudy.

B.

I,, j

/

238 THE JOURNALOF NUCLEARMEDICINE•Vol. 41 •No. 2 •February 2000

. . .. . Cortex

200- •..o.@Cerebellum::@ ••A•Specific@ 150

t 100@@ 50

0@@

0 100 200 300 400 500 600Minutes

by on June 30, 2020. For personal use only. jnm.snmjournals.org Downloaded from

Volunteer no. % Change/h*

% Change/h*

MetabolitesandROls Mean SD

1

Totat/V3V3Cortexcerebellar@ V31t (meanfi)t(mdiv. f1)@

TABLE12-

-C,) 1.5- +-+@ i 1-•@@ ...C.)

0 100 200 300 400 500 600

Minutes

Rates of Change in PlasmaMetabolitesand PET BrainROlsat Equilibrium

PlasmametaboliteAltanserint“FBP―Altanserinol

BrainAOlAnteriorcingulatecortexFrontalcortexOrbitofrontal cortexParietalcortexTemporalcortexOccipitalcortexCombinedcortexCerebellumCorticalspecificbinding

“FBP―= majorlipophilicmetabolitefraction,probablyderivedfrom 4-(p-fluoro-benzoyl)piperidine, as explained in text.

*Valuesare obtainedby dividingthe linear regressionslopebymeanvaluefrom6 to 10.5h afterinjectionin7 volunteersstudiedforat least 620 mm.

tParentcompound.

0.69 ±5.213.64 ±1.905.78±9.02

3.33 ±1.131.38±1.842.70 ±3.112.78 ±1.222.50±1.302.67 ±1.612.46 ±1.033.45±2.101.16±2.00

FIGURE4. Stabilityof outcomemeasureV3' forcombinedcortex (mean ±SEM used for legibility)after [18F]altanserininfusion in 7 volunteers studied for at least 620 mm. (Onevolunteerwhoseinfusiondurationwas555 mmisnotincluded.)ForeachPETacquisitionofeachvolunteer,V3' (ratioofspecificuptake[cortical—cerebeilar]to total plasmaparentconcentration)wasderivedusingplasmasamplenearestintimetomidpointofPETacquisition.From6 honward,meanrateofchangewaslowforcorticalV3'(1.11±1.69%/h).

%Ih for the 7 volunteers studied for at least 620 mn.Moreover, this rate of change did not exceed ±5%/h for anyvolunteer (Table 3).

DISCUSSION

The major finding of this study is that [‘8F]altanserinis asuitable radioligand for bolus plus constant infusion studiesfor equilibrium modeling with PET. Stable time—activitycurves for both plasma and PET measurements suggest thata state of sustained equilibrium is achieved by approximately 6 h, permitting direct measurement of distributionvolumes V3 and V3'. These results afford an alternative tokinetic modeling for quantification of 5-HT2A receptors inthe human brain using [‘8F]altanserinand PET.

The proposed equilibrium model for [‘8F]altanserinsug

TABLE3

and parietal cortex (r —0.69;P < 0.05). However, valuesfor V3 using individual f1 values were uncorrelated with age.This age-related decline in 5-HT2A receptors has beenpreviously reported in both postmortem (5) and in vivo PET(35—37)studies.

Once equilibrium was achieved, values of V3' (andtherefore V3, using mean or individual values of f1) demonstrated excellent stability as shown in Figure 4 and Table 3.From 6 h onward, the rate of change for V3' was 1.11 ±1.69

TABLE2OutcomeMeasuresat Equilibriumwith[18F]Altansermn

in 8 Volunteers*

Antenorcingulate1.97 ±0.161.57 ±0.38169 ±41248 ±149Frontal1.63±0241.02±0.39110±42179±130Orbitofrontal1

.79±0.151 .28±0.35138 ±37204 ±126Parietal1.88±0.131.43±0.34153±37224±135Temporal@l

.74 ±0.181 .20±0.38129 ±41187 ±127Occipital1.70±0.131.17±0.41126±44175±102Combined1

.75 ±0.151 23 ±0.32132 ±34198 ±128

Stabilityof V3' at Equilibriumin IndividualVolunteers

*All valuesare obtainedas the averageof data between6 and10.5hafterinjection.

tV3' @5calculatedas ratioof specific(cortex—cerebellum)to theplasmatotalparentconcentration(C@/C@).V3iscalculatedas ratioofspecificto plasmafree parentconcentration(C@/f1C@;equivalenttoBP).

tV3 (meanf1)usesgroupmeanvalueoff1 (0.93%).§V3(mdiv.f1)usesindividualvaluesoff1.

3.082 2.793 —1.684 —0.215 1.836 1.417 0.59Mean 1.11SD 1.69

*Valuesareobtainedbydividinglinearregressionslopebymeanvaluefrom6to 10.5h afterinjection.

PET QUANTIFICATIONOF 5-HT2A RECEPTORS•van Dyck et al. 239

by on June 30, 2020. For personal use only. jnm.snmjournals.org Downloaded from

gestsa novel methodto overcomedifficultiesassociatedwith lipophilic radiolabeled metabolites. Whereas kineticstudies with [‘8F]altanserin(16—19)have required complexmodeling to identify a large number of rate constants,equilibrium modeling with [‘8F}altanserin(assuming urnform uptake of nondisplaceable [‘8F]altanserinand metabolites) permits the estimation of specific binding in a receptorrich cortical region by simply subtracting cerebellar activity.

PlasmaMeasurementsThe plasma analysis used in these experiments allowed

quantitating parent [‘8F]altanserinseparated from metabolites, but only clearly resolved the reduced alcohol [‘8F]altanserinol. What we have called the “FBP―fraction actuallyconsists of a mxture of metabolites that probably resultfrom secondary metabolism of FBP. Detailed studies by thePET group at the University of Pittsburgh have shown that[‘8F]FBPis converted to at least 1 other metabolite in ratsand baboons (20). They reported that after administration of[‘8FJFBPto baboons, radioactivity was uniformly distributed and was not displaced by the 5-HT2A antagonist SR46349B (21). In this study, the fact that about 60% of[‘8F]altanserinremained unmetabolized in plasma at earlytime points (37 and 67 mn after injection) is consistent withprevious kinetic studies (1 7). This value later stabilized atabout 45% during the period of sustained equilibrium.

PETMeasurementsThe overall regional variability of uptake of [‘8F]altan

serin in this equilibrium study agreed with that of previouskinetic studies (16—19)as well as the known distribution ofS@T2A receptors as seen in autoradiographic studies (31—

33). After intravenous injection of a bolus plus constantinfusion of [‘8F]altanserin,radioactivity was high in allcerebral cortical regions and low in basal ganglia, thalamus,and cerebellum.

Ratios of total-to-nonspecific (cerebellar) binding between 6 and 10.5 h after injection were 1.97 ±0.16 for thestructure of highest uptake (anterior cingulate) and 1.75 ±0.15 for whole cortex. The fact that these ratios weresomewhat lower than previously reported for [‘@F]altanserinkinetic studies (2.6) (17) is likely the result of: the greaterage of our volunteers (mean, 32 versus 25 y); and the lowerratio of unmetabolized [‘8F]altanserininherent in a prolonged equilibrium compared with a kinetic paradigm (45%after 6 h of constant infusion versus 60% at 1 h after bolusinjection). Our group is currently exploring the deuteriumsubstitution ofthe 2'-hydrogen ofaltanserin ([‘8F]deuteroaltanserin), yielding a metabolically more stable radiotracerwith higher ratio of parent tracer to radiometabolites (38).

DistributionVolumesThe values of V3 (as an estimate of the BP) calculated in

this study (Table 2) are on the order of magnitude predictedfrom in vitro data. Pazos et al. (32) reported 5-HT2 receptordensities of approximately 200—270fmol/mg protein forthose cortical regions analyzed in this study. Assuming

approximately 10% protein in wet brain and 50% whitematter in our PET ROIs, we would expect a B,@ on theorder of 10—13.5nmol/L. The previously published K1value(0.13 nmol/L) of altanserin for 5-HT2 receptors (13) wouldthus predict BP values in the range of 77—104for our corticalROIs.

The outcome measure of choice for clinical studiesremains to be determined. The stability in V3 and V3' onceequilibrium is achieved suggests that a single long PETacquisition (e.g., 40 mm) obtained at 6 h may provide areasonable measure of 5-HT2A receptor density. V3 (usingindividual measurements of f1) has the theoretical advantageofdirectly estimating the BP but may lack reliability becauseof the difficulty of measuring f1. The wide intersubjectvariability in f1 points to the need to examine the intrasubjectvariability in f1 as well. V3' (or the proportionate measure ofV3 using a group mean f1) rests on the questionableassumption of negligible intersubject variability in f1, butmay constitute a more reliable measure. Further test—reteststudies are needed to establish the within-subjectreliabilityof both outcome measures obtained from a single PET scan6 h postinjection.

Assumptionsof @8F]AItanserInEquilibriumModelingThe assumptions underlying the present model include:

that cerebellum is relatively devoid of 5-HT@ receptors andthus lacks displaceable uptake; and that the nondisplaceableC2 and C,@ are uniformly distributed throughout the brain.Support for the second assumption is furnished by both ratautoradiographic studies (39) as well as human PET kineticstudies in which ketanserin pretreatment produced nonspecific binding of [‘8F]altanserin and metabolites that wasuniform in all brain regions (17).

With regard to the first assumption, the cerebellum haspreviously been used as a reference region for determinationof S@T2A receptors in PET studies using [“C]NMSP(9,35),[‘8F]setoperone(10,40), and [‘8F]altanserin(16). The choiceof cerebellum as a reference region is based on the lowlevels of S@T2A receptors from postmortem studies (31—33). The aforementioned ketanserin pretreatment studieswith [‘8F]altanserin(17) supported the validity of estimatingthe nonspecific binding in cerebellum, revealing similar[‘8F]altanserinkinetics in cerebellum during blocked andunblocked studies in the same individuals. Further competition studies should be conducted in nonhuman primates withcompetitors of well-defined specificity. The presence of asmall component of specific binding in cerebellum mayresult in the underestimation of specific uptake in receptorrich cortical regions using the proposed equilibrium model.

CONCLUSION

These results demonstrate the feasibility of equilibriumimaging with [‘8F]altanserinand suggest a method toovercome difficulties associated with lipophilic radiolabeledmetabolites. Further studies are needed to verify assumptions of the proposed model and to establish the reliability ofoutcome measures obtained from a single PET scan.

240 THE JOURNALOF NUCLEARMEDICINE•Vol. 41 •No. 2 •February 2000

by on June 30, 2020. For personal use only. jnm.snmjournals.org Downloaded from

19. Smith GS, Price JC, Mathis CA, et al. Test-retest variability ofserotonin 5-HT@

receptorbinding measured with positron emission tomography and [‘@F]a1tanserinin the human brain. Synapse. 1998:30:380—392.

20. Mason NS, Huang Y, Holt DP, Perevuznik JJ, Lopresti Bi, Mathis CA. Synthesisand characterization of [F-18]4-(4-fluorobenzoyl)piperidine, an [F-18]altanserin

metabolite [abstract]. JNucl Med@1997;38:56P.

21. Mason NS, Huang Y, Holt DP. Perevuznik JJ, Lopresti BJ. Mathis CA. Synthesisof two radiolabeled metabolites of [“F]altanserin:[@F]3-[2-[4-(4-fluorophenyl

methanol)-1-piperidinyl]ethyl]-2.3-dihydro-2.thioxo-4-(IH)-quinawlinone and

[‘8F]4-(4-fluorobenzoyl)piperidine.J Labelled Compds Radiopharm. 1997:40:5!161—162.

22. Frey MA, Ehrenkaufer LE, Beaucage 5, Agranoff BW. Quantitative in vivoreceptor binding I. Theory and application of the muscarinic cholinergic receptor.

JNeurosci. 1985:5:421—428.23. Carson RE, Channing MA, BIaSbCrg RG, et al. Comparison ofbolus and infusion

methods for receptor quantitation: applications to [‘8F]cyclofoxyand positron

emission tomography. J Cereb BloodFlow Metab. 1993:13:24—42.24. Tan P-Z, Baldwin 1W, Soufer R, Garg PK, Charney DS. Innis RB. A complete

remote-control system for reliable preparation of [“F]altanserin,radioligand for

imaging of 5-HT2A receptor in brain with positron emission tomography. Appl

Radio: Isot. 1999;50:923—927.

25. Monclus M, Luxen A. A convenient synthesis of 4-nitrophenyl 4-piperidinylketone. Organ Prep Pm In:. l992;24:692-694.

26. USP XXIII. Sterility tests <71>. In: The United States Pharmacopeia. 23rd rev

ad.Rockville,MD: UnitedStatesPharmacopeialConvention;1995:1686—1690.27. Mullani NA, Gould KL, Haitz RK, ci al. Designandperformanceof Posicam6.5

BOO positron camera. JNuclMed@1990;31:6l0-616.28. Matsui T, Hirano A. An Atlas ofthe Human Brainfor ComputerizedTomography.

Tokyo, Japan: Igaku-Shoin, Ltd; 1978.29. Gandelman MS. Baldwin RM, Zoghbi 55, Zea-Ponce Y, Innis RB. Evaluation of

ultrafiltration for the free fraction determination of single photon emissioncomputed tomography (SPECT) tracers: 6-CIT. IBF, and iomazenil. J Pharm Sci.1994:83:1014-1019.

30. Mintun MA, Raichle ME, Kilbourn MR. Wooten GF, Welch MJ. A quantitative

model for the in vivo assessment of drug binding sites with positron emissiontomography.Ann Neurol. 1984:15:217—227.

31. Schotte A, Maloteaux SM. Laduron PM. Characterization and regional distribu

tion ofserotonin Si-receptors in human brain. Brain Res. 1983:276:231—235.

32. Paws A, Probst A, Palacios JM. Serotonin receptors in the human brain. IV.Autoradiographic mapping of serotonin-2 receptors. Neuroscience. 1987:21:123—139.

33. De Keyser J, Clayes A, Dc Backer JP, Ebinger G. Rods F, Vauquelin G.

Autoradiographic localization of Dl and D2 receptors in the human brain.

NeurosciLen.1988;91:142—147.34. LaruelleM, BaldwinRM, RaunerZ, et al. SPECTquantificationof['@fliomaze

nil binding to benzodiazepine receptors in nonhuman primates. I. Kineticmodeling of single bolus experiments. J Cereb Blood Flow Metab. 1994:14:439—

452.35. Wong D, Lever J, Hartig P. et al Localization of serotonin 5-HT2 receptors in

living human brain by positron emission tomography using NL([hlC]@methyl)@2@Br-LSD. Synapse. 1987; 1:393—398.

36. RosierA, Dupont P. Peuskens3, et al. Visualisationof lossof 5-HT@ receptorswith age in healthy volunteers using [“F]altanserin and positron emission

tomographic imaging. Psychiatry Res Neumimaging. 1996:68:11—22.37. Meltzer CC, Smith G, Price IC, et al. Reduced binding of [“F]altanserinto

serotonin type 2A receptors in aging: persistence of effect after partial volume

correction. Brain Res. l998;813:167—l71.38. Tan P-Z, Baldwin RM, Fu T, Charney DS, Innis RB. Rapid synthesis of F-l8 and

H-2 dual-labeled altanserin, a metabolically resistant PET ligand for 5-HT2Areceptors. JLabelled Compds Radiopharnt 1999:42:457-467.

39. Biver F,Goldman5, MonclusM, LuxenA, LotstraF,MendlewiczJ. Autoradiographic study of [‘8F]altanserinin the rat: a potential radiotracer for 5HT2receptorstudy with PET. Clin Neuropharmacol. l992;l5(suppl l):202.

40. Blin 3, BaronJ, DuboisB, et al. Lossof brain 5-HT2 receptorsin Alzheimer'sdisease. Brain. 1993:116:497—510.

ACKNOWLEDGMENTS

The authors thank Dayton Rich and Quinn Ramsby forexcellent technical assistance and Dr. Yolanda Zea-Ponce forhelping with initial metabolite analysis method development. This research was supported by funds from theDepartment of Veterans Affairs (Schizophrenia ResearchCenter and the VA Mental Illness Research, Education, andClinical Center) and the Public Health Service (MH 30929).

REFERENCESI. Arango V. Ernsberg P. Mazuk PM, et al. Autoradiographic demonstration of

increased 5-HT2 and 8-adrenergic receptors in the brains of suicide victims. ArchGenPsvchiatrj.1990;47:1038—1047.

2. Stanley M, Mann ii. Increased serotonin-2binding sites in frontal cortex ofsuicide victims. Lance:. 1983;2:214—216.

3. Mita 1, Hanada S. Nishino N, et al. Decreased serotonin S2 and increaseddopamine D2 receptors in chronic schizophrenics. BiolPsychia:r.y. 1986;21:1407—1414.

4. Laruelle M, Abi-Dargham A, Casanova MF. lou R. Weinberger DR. Kleinman JE.

Selective abnormalities of prefrontal serotonergic receptors in schizophrenia: apostmortem study. Arch Gen Psychiatry. 1993;50:810—818.

5. MarcussonJO, Morgan DO, Winbiad B. Finch CE. Serotonin-2binding sitesinhuman frontal cortex and hippocampus: selective loss of S-2A sites with age.Brain Res. 1984;311:51—56.

6. ReynoldsOP.ArnoldL, RossorMN, IversenLL, MountjoyCQ,RothM. Reducedbinding of [3H)ketanserin to cortical 5-HT2 receptors in senile dementia of theAlzheimer type. Neurosci Len. 1984;44:47—51.

7. CrossAJ, CrowTJ, FerrierIN, JohnsonJA.The selectivityof thereductionofserotonin S2 receptors changes in Alzhcimer-typc dementia. Neurobiol Aging.1986;7:3—7.

8. Pike VP. Radioligands for PET studies of central 5-HT receptors and re-uptakesites—current status. Nuci MedBioL 1995:22:1011—1018.

9. Wong DF, Wagner MN. Jr, Dannals RE, et al. Effects of age on dopamineandserotonin receptors measured by positron emission tomography in the livinghuman brain. Science. 1984:226:1393—1396.

10. Bun J, Sette G, Fiorelli M, et al. A method for the in vivo investigation of theserotonergic 5-HT2 receptors in the human cerebral cortex using positronemission tomography and ‘@F-labe1edsetoperone. J Neurochem. 1990;54:1744—1754.

11. Lemaire C, Cantineau R, Guillaume M. Plenevaux A, Christians L Fluorine-18-

altanserin: a radioligand for the study of serotonin receptors with PET: radiolabeling and in vivo biologic behavior in rats. JNucIMed. 199l;32:2266—2272.

12. Lundkvist C, Halldin C, Ginovart N, et al. [“C]MDL 100907, a radioligand for

selective imaging of 5-HT@ receptors with positron emission tomography. LifeSci. 1996;58:187—192.

13. Leysen JE. Use of5-HT receptor agonists and antagonists for the characterization

of their respective receptor sites. In: Boulton AB, Baker GB. Juno AV, eds. Drugsas Toolsin NeurotransmiuerResearch.Neummethods.Vol 12. Clifton, NJ: TheHumana Press; 1989:299—349.

14. Mathis C, Choi Y, Simpson N, Mintun M. In vivo studies of 5-HT2 receptors in

monkeys using F-18 altanserin [abstract]. Soc NeumsciAbst. l994;20:l546.15. Tan P-Z, Baldwin RM, van Dyck CH, et a]. Characterizationof radioactive

metabolites of 5-HT2A receptor PET ligand (F-l8jaltanserin in human and rodent.NuciMed Biol. 1999;26:601-608.

16. Biver F, GoldmanS. Luxen A, et al. Multicompartmentalstudyof fluorine-l8altanserin binding to brain 5HT2 receptors in humans using positron emissiontomography. Eur J Nuci Med@l994;21 :937—946.

17. Sadzot B, Lemaire C, Maquet P. et al. Serotonin 5HT2 receptor imaging in thehuman brain using positron emission tomography and a new radioligand,[“F]altanserin:results in young nonnal controls. J Cereb Blood Flow Metab.1995;1 5:787—797.

18. Mintun M, Price J, Smith G, Ctal. Quantitative 5-HT2A receptor imaging in manusing F-18 altanserin: a new model accounting for labeled metabolites [abstract].J NuciMed. l996;37:109P.

PET QUANTIFICATIONOF S@'T2A RECEPTORS•van Dyck et al. 241

by on June 30, 2020. For personal use only. jnm.snmjournals.org Downloaded from

2000;41:234-241.J Nucl Med.   Soufer, Dennis S. Charney and Robert B. InnisChristopher H. van Dyck, Ping-Zhong Tan, Ronald M. Baldwin, Louis A. Amici, Pradeep K. Garg, Chin K. Ng, Robert 

F]Altanserin18Paradigm with [ Receptors in the Human Brain: A Constant Infusion2APET Quantification of 5-HT

http://jnm.snmjournals.org/content/41/2/234This article and updated information are available at:

  http://jnm.snmjournals.org/site/subscriptions/online.xhtml

Information about subscriptions to JNM can be found at:  

http://jnm.snmjournals.org/site/misc/permission.xhtmlInformation about reproducing figures, tables, or other portions of this article can be found online at:

(Print ISSN: 0161-5505, Online ISSN: 2159-662X)1850 Samuel Morse Drive, Reston, VA 20190.SNMMI | Society of Nuclear Medicine and Molecular Imaging

is published monthly.The Journal of Nuclear Medicine

© Copyright 2000 SNMMI; all rights reserved.

by on June 30, 2020. For personal use only. jnm.snmjournals.org Downloaded from