Brain Serotonin Transporter Binding Potential Measured With Carbon11–Labeled DASB Positron Emission Tomography: Title and subTitle BreakEffects of Major Depressive Episodes and Severity of DysfunctionalAttitudes FREE

Serotonin (5-HT) may be abnormally regulated during major depressiveepisodes (MDE) due to major depressive disorder. This assertion is based largelyon abnormalities of serotonin turnover during MDE. For example, the levelof the cerebrospinal fluid metabolite of serotonin, 5-hydroxyindoleaceticacid, is often low during MDE, especially when suicidal ideation is present.^{1 - 2} Prolactin release is increased afteradministration of fenfluramine, a 5-HT–releasing drug, and it has beenobserved that the prolactin release after fenfluramine administration is attenuatedduring MDE.^{3 - 4} A role for serotoninto modulate mood has been proposed, because mood lowering after tryptophandepletion is often observed in subjects with either a family history of depressiveepisodes or a history of depressive episodes.^{5 - 6}

The primary mechanism by which extracellular levels of serotonin maybe low during an MDE due to major depressive disorder is unknown. It has beenproposed that increased serotonin transporter (5-HTT) density could lead toincreased serotonin clearance from extracellular regions in the brain. Theanswer to this fundamentally important question is not known because thereare no investigations of 5-HTT density in subjects who are in the midst ofa depressive episode due to major depressive disorder. Instead, a number ofpostmortem studies have investigated subjects with a history of an MDE. Thesestudies usually do not distinguish between unipolar disorder and bipolar disorder,and the samples of unipolar subjects are often small. Most of these postmorteminvestigations report decreased 5-HTT density^{7 - 9} orno difference in 5-HTT density^{10 - 12} betweendepressed and healthy subjects. Although the regions investigated varied acrossthese studies, all sampled the prefrontal cortex.

It is difficult for postmortem investigations to determine the relationshipbetween 5-HTT density abnormalities and the presence of a current depressiveepisode, because the clinical data are gathered retrospectively. In theory,with receptor-ligand imaging in vivo, it is quite feasible to collect clinicaldata and determine the binding potential (BP), an index of receptor density,during a current depressive episode.

In the past, the barrier to investigating the 5-HTT BP with receptor-ligandimaging was the absence of a method that was both valid for multiple brainregions and reliable. Quantification of the 5-HTT BP in most brain regions,including the prefrontal cortex, was not possible with either of the 2 previousimaging methods, including carbon 11 [¹¹C]–labeled(+)McN5652positron emission tomography (PET) and iodine I 123–labeled 2β-carbomethoxy-3β-(4-¹²³I-iodophenyl)-tropane (CIT) single-photon emission computed tomography(SPECT). To our knowledge, the test-retest reliability for regional 5-HTTBP found with [¹²³I]β-CIT SPECT or [¹¹C](+)-McN5652PET has not been published. Relative to specific binding, [¹¹C](+)McN5652had very high nonspecific uptake, and 5-HTT BP values were only detectablein the thalamus¹³ and possibly the basal gangliaand midbrain (although measuring 5-HTT BP in the latter 2 regions may requireconcurrent arterial sampling of the radiotracer).¹⁴ TheSPECT radiotracer, [¹²³I]β-CIT, has been used to measure 5-HTTsites in the midbrain region in depressed subjects.^{15 - 16} However,[¹²³I]β-CIT has similar affinity for the dopamine transporter(DAT) and the 5-HTT^{17 - 18} ; hencethe midbrain [¹²³I]β-CIT is a combined measure of DAT BP inthe substantia nigra and 5-HTT BP in the raphe nuclei.

Ichimiya et al¹⁹ used [¹¹C](+)McN5652PET to investigate the thalamus 5-HTT BP in 7 subjects with major depressivedisorder. The data from these 7 subjects were pooled with those of 6 subjectswho had bipolar disorder. That study did not address whether 5-HTT BP wasabnormal during a current MDE. Only 5 subjects with a current MDE and majordepressive disorder were enrolled in the study.¹⁹

Carbon 11–labeled 3-amino-4-(2-dimethylaminomethyl-phenylsulfanyl)-benzonitrile(DASB) positron emission tomography (PET) is a new advance in brain imagingthat measures 5-HTT BP in multiple brain regions, including the prefrontalcortex.^{20 - 24} [¹¹C]DASB is highly selective, showing nanomolar affinity for the 5-HTTand negligible affinity for other receptors.^{20 - 21} Themain advantage of [¹¹C]DASB is that it has a much higher ratioof specific binding to nonspecific binding in vivo. As a result, with [¹¹C]DASB PET, 5-HTT BP values are quantifiable in prefrontal cortexand reasonably high in the basal ganglia and thalamus, even without arterialsampling.^{22 - 23} Furthermore, the5-HTT BP found with [¹¹C]DASB PET is reliable for most brain regionsunder test-retest conditions.²⁴

This study had 2 main purposes. The first was to investigate the prefrontaland subcortical (thalamus, basal ganglia, and midbrain) 5-HTT BP in subjectswho are currently experiencing an MDE secondary to major depressive disorder.The first purpose focused on whether brain 5-HTT abnormalities occur duringa current MDE, an issue not resolved in earlier postmortem studies^{8 - 12} orthe PET investigation of thalamic 5-HTT BP by Ichimiya et al.¹⁹

The second main purpose was to determine the relationship between 5-HTTBP and dysfunctional attitudes in MDE. Dysfunctional attitudes are negativelybiased assumptions and beliefs regarding oneself, the world, and the future.A modest degree of negativism is found in healthy subjects, whereas a highproportion of subjects with MDE have severely negativistic thinking. Our recentinvestigations suggest that a subgroup of subjects with MDE, rather than allsubjects with MDE, have very low levels of extracellular serotonin. A previousstudy from our group found that increasing 5-HT agonism with a single doseof d-fenfluramine in healthy subjects lowered dysfunctional attitudes tremendouslytoward optimism (highly significant effect, F_1,25 = 17[P < .001]).²⁵ Thisfinding suggested a role for serotonin in humans as a modulator of dysfunctionalattitudes. Given that serotonin can function as a modulator of dysfunctionalattitudes, it seemed quite possible that low levels of serotonin could contributeto the increased severity of dysfunctional attitudes that often occur duringMDE. In a second investigation in MDE subjects, we found that negativisticdysfunctional attitudes were associated with increased cortex serotonin₂ BP.²⁵ Furthermore, MDE subjects withseverely negativistic dysfunctional attitudes had increased serotonin₂ BP compared with healthy subjects. Both findings were highly significant,especially in the prefrontal cortex (r = 0.56[P = .009] and F_1,19 = 11[P = .003], respectively).²⁵ Inanimal models, serotonin₂ receptor density may increase after certain5-HT–depleting paradigms (that reduce serotonin synthesis or storagefor ≥2 weeks^{26 - 27} ) and decreaseafter certain 5-HT–increasing paradigms (monoamine oxidase inhibitionfor ≥2 weeks^{28 - 29} ). The BPis proportional to receptor density. A plausible explanation for the relationshipbetween serotonin₂ BP and dysfunctional attitudes is that subjectswho have a form of MDE with severely negativistic dysfunctional attitudesalso have had an extended period of low levels of extracellular serotoninin the cerebral cortex.

We hypothesized that MDE subjects with greater 5-HTT BP would have moreseverely negativistic dysfunctional attitudes. The underlying model for thishypothesis is that MDE subjects with greater 5-HTT BP will remove more extracellularserotonin, have lower levels of extracellular serotonin, and experience moresevere symptoms of pessimism. The prefrontal cortex region is the primarylocation for this hypothesis, because it is typically sampled in investigationsreporting increased serotonin₂ density in drug-free depressed suicidevictims.^{12 ,30} Furthermore, thisregion had the strongest correlation between the Dysfunctional Attitudes Scale(DAS) and serotonin₂ BP in a previous study by our group.²⁵ We also examined the relationship between dysfunctionalattitudes and other sampled brain regions, including the midbrain. However,the 5-HTT in the midbrain region is proximal to serotonergic cell bodies³¹ and would not be expected to directly modulate serotoninlevels in distant brain regions.

Twenty subjects with an MDE and major depressive disorder (mean age,35 years [SD, 11 years]; 9 men and 11 women) and another 20 age-matchedhealthy subjects (mean age, 35 years [SD, 11 years]; 10 men and 10 women)were recruited. Ages ranged from 19 to 52 years. Healthy subjects were agematched within 3 years to depressed patients. All MDE and healthy subjectswere physically healthy, had no history of alcohol or substance abuse or neurotoxinuse, were free of psychotropic drug use for longer than 3 months plus 5 half-livesof any medication, and were nonsmoking.

Healthy subjects underwent screening to rule out Axis I disorders (currentor in remission),³² current suicidal ideation,and history of self-harm behavior, anger dyscontrol, or impulsive behavior.For each subject, written consent was obtained after the procedures had beenfully explained.

The 20 subjects with MDE were obtained from a larger sample of 37 medication-freesubjects with MDE who had enrolled in a study of antidepressant occupancy.^{24 ,33} All were followed up for a minimumof 2 months by a psychiatrist (J.H.M.). The other 17 subjects were excludedon the basis of the criteria of nonsmoking and comorbid Axis I anxiety disorders.Diagnosis of MDE secondary to major depressive disorder was based on the StructuredClinical Interview for DSM-IV for Axis I disorders(SCID), (S.S.)³⁴ and a consultation by a psychiatrist(J.H.M). For subjects with MDE, the minimum severity of depression for enrollmentwas based on a cutoff score of 16 on the 17-item Hamilton Depression RatingScale. The mean Hamilton Depression Rating Scale score was 20 (SD, 4). Theserecruiting methods are similar to what has been described previously.^{25 ,35} Exclusion criteria included MDE withpsychotic symptoms, bipolar disorder (type I or II), comorbid Axis I diagnoses,and history of self-harm or suicidality outside of episodes of depression,anger dyscontrol, impulsive behavior, and neuroleptic use. The exclusion ofself-harm, anger dyscontrol, and impulsivity (taken from the Structured ClinicalInterview for DSM-IV Axis II Disorders)³⁶ ruledout severe borderline personality disorder behaviors. People who have severeborderline personality disorder behaviors have abnormal severity of dysfunctionalattitudes.³⁷ Fourteen of the 20 depressed patientshad never received a trial of antidepressant treatment. No subject with MDEhad received antidepressant treatment within the past 3 months. The recruitmentof medication-free subjects has become a recognized standard in receptor-ligandimaging studies of MDE.^{25 ,35 ,38 - 40}

All patients received common blood tests to rule out medical causesof disturbed mood (measurement of thyroid function, electrolyte levels, andcomplete blood cell count). Three of the MDE subjects who had received anantidepressant trial in the past also had a history of substance use (marijuana)that did not meet criteria for the SCID diagnosis of drug abuse. These 3 patientshad not used marijuana within the previous year. Each underwent urine drugscreening, the results of which were negative. All healthy subjects underwenturine drug screening.

All patients gave written consent after the procedure had been fullyexplained. The study and recruitment procedures were approved by the researchethics board for human subjects at the Centre for Addiction and Mental Health,Toronto, Ontario.

Additional information recorded included version A of the DAS; visualanalog scales for mood, anxiety, and energy; and the Scale for Suicide Ideation.⁴¹ Version A of the DAS is a 40-item self-report thatrequires subjects to indicate to what degree they agree with an individualstatement. Each endorsement is converted into a score between 1 and 7. Thetotal score for endorsement of pessimistic statements and disagreement withoptimistic statements becomes the final measurement from the DAS. The DASdetects negativistic thinking during episodes of MDE,^{42 - 43} hasvery good test-retest reliability across 6 to 8 weeks (r = 0.83-0.84),^{44 - 45} andhas a high degree of internal consistency (Cronbach α = 0.85-0.87).^{45 - 46}

The synthesis and imaging of [¹¹C]DASB has been describedpreviously.^{20 - 23} Beforethe PET scan, an intravenous bolus of 10 mCi (370 MBq) of [¹¹C]DASBwas injected. The [¹¹C]DASB was of high radiochemical purity (>95%)and high specific activity (950 ± 270 mCi/μmol [35 ± 10GBq/μmol] at the time of injection). The PET images were obtained usinga GEMS 2048-15B camera (Scanditronix Medical, General Electric, Uppsala, Sweden;intrinsic resolution, full width at half maximum, 5.5 mm). The images werecorrected for attenuation using a germanium Ge 68 transmission scan and reconstructedby filtered back projection (Hanning filter).

To obtain a measure of the 5-HTT BP with region-of-interest data, wechose the noninvasive method of Logan et al⁴⁷ (hereafterreferred to as the Logan method) implemented within PMOD software (PMOD TechnologiesLtd, Zurich, Switzerland).⁴⁸

We used the noninvasive Logan method because it provides valid measurementsof 5-HTT BP, and the between-subject variance in 5-HTT BP is low for mostbrain regions. The coefficient of variation (standard deviation/mean) usingthe Logan method⁴⁷ varied from 25% in the prefrontalcortex and midbrain to 15% in the putamen. Furthermore, the 5-HTT BP foundwith this method is very reproducible. In our data sets,²⁴ theabsolute difference in the test-retest measurement of the 5-HTT BP with applicationof the Logan method⁴⁷ averages below 15% ofthe 5-HTT BP itself in all regions except the dorsolateral prefrontal cortexand midbrain, for which it is higher at about 15% to 20% of the mean 5-HTTBP. The less reliable test-retest measurement of 5-HTT BP in the dorsolateralprefrontal cortex region and midbrain reflects a lesser sensitivity of [¹¹C]DASB to detect the 5-HTT BP in these regions with this method. Themeasurement of 5-HTT BP with the noninvasive Logan method⁴⁷ isan underestimate of the same 5-HTT BP found with invasive arterial samplingand the single-tissue compartment model. This is quite acceptable becausethe between-subject variance of 5-HTT BP is also lower with the noninvasiveLogan method,⁴⁷ and the regional 5-HTT BP foundwith the 2 methods is very strongly correlated. In general, we find that the5-HTT BP measured using reference tissue–based methods is highly correlatedwith the ratio of the kinetically determined distribution volumes betweenregions with specific binding and the reference region.²³ Aspart of our assessment of the noninvasive Logan method,⁴⁷ wedetermined distribution volumes in the thalamus, striatum, frontal cortex,and cerebellum in a set of healthy subjects using an arterial input functionand a single-tissue compartment model. Ratios of distribution volumes in regionswith specific binding to the distribution volume of the cerebellum were determinedfor each brain region, and this ratio was highly correlated with the bindingpotential found with the noninvasive Logan method (r = 0.97).

For each region, except the midbrain, analyses were also carried outusing the modified simplified reference tissue method,⁴⁹ whichis also suitable for [¹¹C]DASB PET.⁵⁰ Thelesser reversibility of radiotracer uptake in the midbrain precluded routineuse of the modified simplified reference tissue model in this region.

A key assumption in the noninvasive models is that there is a referenceregion that does not contain specifically bound radioligand. For [¹¹C]DASB, the cerebellum is a suitable reference region because studiesreport either undetectable⁵¹ or extremely low5-HTT density.^{52 - 53} More recently,it was found that the 5-HTT density in the cerebellum is less than 3% of thestriatal value using the Western blot method (Stephen J. Kish, PhD, YoshiakiFurukawa, MD, Li-Jan Chang, MSc, Junchao Tong, PhD, N.G., A.A.W., S.H., andJ.H.M., unpublished data, May 2004).

The BP found with these noninvasive methods represents the ratio ofspecifically bound radiotracer to radiotracer in free and nonspecific compartmentsat equilibrium. This BP is equal to (1/V₂) × B_max/K_d^{54 - 55} where V₂ represents free and nonspecific binding at equilibrium; B_max, receptor density; and K_d, the dissociation constant. For [¹¹C]DASB PET, V₂ is sufficiently similar among subjects that it may beconsidered a constant.²³

Regions chosen were bilateral anteromedial prefrontal cortex (Brodmannareas included were part of 8, 9, and 10), left and right dorsolateral prefrontalcortex (part of Brodmann areas 9 and 46), anterior cingulate, left and rightcaudate, left and right putamen, left and right thalamus, and midbrain. Prefrontalregions were included because indexes of increased serotonin₂ receptordensity have been reported in prefrontal cortex in suicide victims^{12 ,30 ,56 - 59} andmedication-free depressed subjects who are suicidal^{12 ,30} orhave severely negativistic thinking.²⁵ Theanterior cingulate was chosen because metabolic abnormalities have been reportedin this general region in subgroups of depressed subjects.^{60 - 61} Otherregions (thalamus, basal ganglia, and midbrain) were chosen because the 5-HTTdensity is high in these regions.^{51 - 53}

To assist with the region-of-interest measurement, each subject underwenta magnetic resonance imaging scan (GE Signa 1.5-T scanner; GE Medical Systems,Milwaukee, Wis) (spin-echo sequence proton density–weighted image; x,y, and z voxel dimensions, 0.78, 0.78, and 3 mm, respectively). Regions ofinterest were found using a semiautomated method^{62 - 63} verifiedby visual assessment with reference to a coregistered magnetic resonance imagingscan. The only exception was the midbrain region, which was drawn within thesuperior 2 planes to the pons with reference to the coregistered magneticresonance imaging scan. These methods have been described in more detail previously.⁶²

There was no effect of age on regional 5-HTT BP in the healthy samplefor any region (prefrontal cortex, anterior cingulate, bilateral caudate,bilateral putamen, bilateral thalamus, and midbrain) (1-way analysis of variance[ANOVA], F_1,18 = 0.01-3.0 [P = .92-.10]).There was no effect of sex on 5-HTT BP in these regions within the healthysample (1-way ANOVA, F_1,18 = 0.001-1.0 [P = .97-.32]). The results for this analysis in peripheralregions using the modified simplified reference tissue model^{49 - 50} weresimilar (effect of age, 1-way ANOVA, F_1,18 = 0.03-1.8[P = .73-.20]; effect of sex, 1-way ANOVA,F_1,18 = 0.008-0.6 [P = .93-.45]).

In the entire sample, MDE had no effect on regional 5-HTT BP for anyindividual region (1-way ANOVA, F_1,38 = 0.09-1.4 [P = .77-.24]); in fact, the mean regional 5-HTTBP within each group was very similar (Figure1). The results for this analysis in peripheral regions using themodified simplified reference tissue model^{49 - 50} weresimilar (effect of depression, 1-way ANOVA, F_1,38 = 0.1-1.3[P = .75-.26]).

In MDE subjects, in all regions primarily sampling 5-HT nerve terminals(ie, everywhere except the midbrain), there were strong correlations betweenincreasing DAS scores and increasing 5-HTT BP (Figure 2). In healthy subjects, the correlations between DAS andregional 5-HTT BP were nonsignificant. The correlation coefficients betweenthe DAS and regional 5-HTT BP in healthy subjects were compared with the correlationcoefficients between the DAS and regional 5-HTT BP found in MDE subjects.We performed the comparison by transforming the correlation coefficients into z scores with the Fisher transformation so that they thencould be compared using the function of the gaussian distribution.⁶⁴ For all regions mainly sampling 5-HT nerve terminals,the correlation between regional 5-HTT BP and DAS was significantly differentbetween MDE and healthy subjects (Table).For all of these analyses, extremely similar results were obtained using regional5-HTT BP found with the modified simplified reference tissue method.^{49 - 50}

Severely negativistic dysfunctional attitudes in the MDE subjects weredefined as a DAS score of greater than 190. The cutoff score of 190 was chosenbecause this score is 3 SDs above the mean DAS score of our healthy subjects.Eight subjects with MDE had a DAS score greater than 190, and their regional5-HTT BP was compared with that of the 20 healthy subjects. For all peripheralregions, the 5-HTT BP was higher in the MDE subjects with highly abnormalDAS scores (Figure 3). For these analyses,similar results were obtained using regional 5-HTT BP found with the modifiedsimplified reference tissue method.^{49 - 50}

To our knowledge, this was the first investigation of regional 5-HTTBP with a selective radioligand in drug-free depressed subjects, and therewere 3 main findings. First, there was no difference in regional 5-HTT BPin the entire sample of depressed subjects compared with the healthy subjects.Second, depressed subjects with greater regional 5-HTT BP had higher levelsof dysfunctional attitudes. Third, a subgroup of depressed subjects with extremelyhigh levels of dysfunctional attitudes had increased regional 5-HTT BP comparedwith healthy subjects.

An important function of the 5-HTT is to remove extracellular serotonin.This concept is primarily based on the findings that selective serotonin reuptakeinhibitors inhibit the reuptake of 5-HT⁶⁵ andthat knockout mice without the 5-HTT have increased levels of extracellularserotonin.⁶⁶ The findings of greater 5-HTTBP in MDE subjects with high levels of dysfunctional attitudes is consistentwith this role. Greater 5-HTT BP could result in greater clearance of serotonin.On the basis of our previous findings,²⁵ theMDE subjects with more negativistic thinking should have lower levels of extracellularserotonin. (Previous findings were that increased serotonin release afterd-fenfluramine administration decreased dysfunctional attitudes toward optimism.²⁵ Higher serotonin₂ BP was found in MDEwith severe dysfunctional attitudes.²⁵ Otherinvestigators have reported greater serotonin₂ density after 5-HT–depletingparadigms of ≥2 weeks’ duration^{26 - 27} andreduced serotonin₂ density after certain serotonin-increasing paradigmsof ≥2 weeks’ duration^{28 - 29} ).The association between negativistic thinking and higher 5-HTT BP is interpretedto reflect a vulnerability to greater extracellular clearance of serotoninwhen the 5-HTT BP is high. Greater extracellular clearance of serotonin thenresults in more severe symptoms. The earlier investigations²⁵ suggestthat depressed patients with greater dysfunctional attitudes have a greaterloss of extracellular serotonin, and the present study proposes that greater5-HTT BP is a mechanism that could contribute to greater extracellular serotoninloss.

This interpretation suggests that the 5-HTT and DAT have similar rolesof increasing vulnerability to symptoms during MDE. In a previous report ofa separate sample of subjects, we found that MDE subjects with the lowestDAT BP were not motor retarded, whereas MDE subjects with higher DAT BP wereimpaired on neuropsychological measurement of motor speed.³⁵ Thecorrelation between DAT BP and motor retardation was very strong (r = 0.86 [P = .006]),and the subjects were also carefully recruited (drug free, nonsmoking, andno comorbidity).³⁵ The interpretation of thisprevious study was that greater DAT BP provides vulnerability to greater clearanceof dopamine, resulting in lower levels of extracellular dopamine and moresevere motor impairment, whereas lower DAT BP reduces clearance of extracellulardopamine, resulting in near-normal extracellular dopamine levels and negligiblemotor impairment.

We are characterizing the sample of MDE subjects with more severe dysfunctionalattitudes as having more severe serotonergic disturbances with greater serotonin₂ BP in the previous sample,²⁵ and withgreater 5-HTT BP in the current sample. This may have implications for suiciderisk due to MDE. The DAS is highly correlated with the Beck Hopelessness Scaleduring MDE,^{67 - 68} and greater scoreson the Beck Hopelessness Scale are associated with higher likelihood of suicide.^{69 - 70} Therefore, it is possible that duringMDE, in some people, elevated 5-HTT BP may lead to lower serotonin levels,less stimulation via major serotonergic signal transduction pathways, andelevated serotonin₂ BP. It may be that these processes mediategreater negativistic thinking and eventually result in an elevated risk ofsuicide.

The correlations between DAS and 5-HTT BP were very strong in the brainregions that primarily sampled serotonergic nerve terminals. This is not surprising,because the 5-HTT BP tends to associate across these regions within subjects.For example, subjects with higher 5-HTT BP in the frontal cortex tend to havehigher 5-HTT BP in the striatum.

In a large earlier postmortem study ( 53 patients vs 107 subjectswith no history of a depressive episode), it was found that the 5-HTT densitywas lower in the patient sample.^{7 - 9} Thereare several major differences in sampling between the earlier study and thepresent one. In our study, all subjects were in the midst of a current MDEdue to major depressive disorder. The previous study sampled patients whohad a history of unipolar or bipolar depressive episodes and were not necessarilyin the midst of a depressive episode at the time of measurement. Another keydifference was that the subjects in the present study were about a decadeyounger than the subjects in the postmortem study. If there are any long-termeffects of mood disorder illnesses on 5-HTT density, the results of the studieswould be expected to differ. Future longitudinal studies of 5-HTT BP in untreatedremitted depressed subjects may help resolve these questions.

Whenever multiple regions are investigated in a brain imaging study,one should consider how multiple comparisons influence the interpretationof results. In this study, the 2 major findings (correlation between 5-HTTBP and dysfunctional attitudes during depression and higher 5-HTT BP in depressedsubjects with severe dysfunctional attitudes) were present in all 6 regionsof serotonin nerve terminal areas. The probability of having significant findingsin all 6 brain regions sampling serotonergic nerve terminals by independentchance alone is exceedingly small. The probability of 6 significant, independentfindings is 0.05⁶ or 1.6 × 10⁻⁸. Therefore, it is unlikely that the findings of this studyrepresent independent chance events.

The results of this study show strong significant correlations betweenthe 5-HTT BP and dysfunctional attitudes. However, there are limitations.Not all brain regions were analyzed; rather, we chose brain regions in whichthe 5-HTT BP could be reliably and validly measured. The 5-HTT BP is proportionalto both 5-HTT density and affinity. Even so, the combined measurement of densityand affinity obtained with the BP does not alter our interpretations, becauseboth 5-HTT density and affinity should contribute to serotonin clearance.We also acknowledge that 5-HTT density and affinity are not the only indicesof 5-HTT function. For example, it is possible for desensitization processesto occur for some receptors without changing either the density or the affinityof receptors. We cannot measure serotonin directly in the human brain, andin our interpretations we made inferences about extracellular serotonin basedon the strong relationships between manipulations of serotonin and dysfunctionalattitudes, as well as measures of serotonin₂ receptors and dysfunctionalattitudes in our previous investigations.²⁵ Wealso recognize that there are alternative interpretations for the correlationsbetween dysfunctional attitudes and regional 5-HTT BP. For example, it couldbe that in MDE with high dysfunctional attitudes, there is a greater densityof neurons expressing 5-HTT, a greater density of dendrites expressing 5-HTT,a wider range of neurons expressing 5-HTT, an unusual regulation of 5-HTTin response to some other process, etc. The importance of the explanationwe presented is that it most straightforwardly accounts for both existingknowledge regarding 5-HTT function^{65 - 66} andthe relationships between dysfunctional attitudes and serotonin measures inearlier investigations.²⁵

This study was the first investigation of the 5-HTT BP in drug-free,nonsmoking, currently depressed subjects with a reasonably large sample sizeusing a specific radioligand. We found that MDE subjects with more negativisticthinking had greater 5-HTT BP in MDE and that a subgroup of MDE subjects withseverely negativistic thinking had a 21% higher 5-HTT BP compared with healthysubjects. There was no difference in 5-HTT BP between the entire sample ofMDE and the healthy subjects. The findings are important because they suggesta role for 5-HTT in the pathophysiology of serotonin-related symptoms. Duringan MDE, elevated 5-HTT BP may result in excessive extracellular clearanceof serotonin, providing vulnerability to serotonin-related symptoms. Thisinterpretation is consistent with previous investigations that associate serotoninagonist effects (ie, d-fenfluramine effect and serotonin₂ BP) withdysfunctional attitudes.²⁵ Our data argue thatthe 5-HTT BP is strongly associated with a symptom of excessive dysfunctionalattitudes within an MDE rather than the MDE itself.

Correspondence: Jeffrey H. Meyer, MD, PhD,FRCPC, 11th Floor, Centre for Addiction and Mental Health, Clarke Division,250 College St, Toronto, Ontario, CanadaM5T 1R8 (jeff.meyer@camhpet.ca).

Submitted for Publication: March 20, 2003;final revision received December 10, 2003; accepted June 23, 2004.

Funding/Support: This study was supported bythe National Alliance for Research in Schizophrenia and Depression (GreatNeck, NY) and Eli-Lilly Canada (Toronto, Ontario). Dr Meyer is supported bythe Canadian Institutes of Health Research New Investigator program (Ottawa,Ontario).

Acknowledgment: We thank research assistantsAlex Kecojevic and Corey Jones and technicians Kevin Cheung, Armando Garcia,Li Jin, MSc, and Ruiping Guo.