- Email: [email protected]
Serotonin transporters in obsessive-compulsive disorder: a positron emission tomography study with [11C]McN 5652
Serotonin Transporters in Obsessive-Compulsive Disorder: A Positron Emission Tomography Study with [11C]McN 5652 H. Blair Simpson, Ilise Lombardo, Mark Slifstein, Henry Yiyun Huang, Dah-Ren Hwang, Anissa Abi-Dargham, Michael R. Liebowitz, and Marc Laruelle Background: Serotonergic abnormalities have been hypothesized to contribute to obsessive– compulsive disorder (OCD). This study examined whether brain serotonin transporter (SERT) availability is altered in OCD using positron emission tomography (PET) and the SERT PET radiotracer [11C]McN 5652. Methods: Eleven OCD subjects, free of psychiatric medications and comorbid depression, and 11 matched healthy control subjects underwent PET scans following injection of [11C]McN 5652 and magnetic resonance imaging (MRI) scans. Total distribution volumes (VT) were derived by kinetic analysis (one tissue compartment model) using the arterial input function. Two measures of SERT availability were computed: binding potential (BP) and specific to nonspecific partition coefficient (V3⬙). Groups were compared using region of interest (ROI) analysis and voxelwise analysis of spatially normalized parametric maps; ROIs were selected based on their relatively high SERT density and included subcortical (dorsal caudate, dorsal putamen, ventral striatum, midbrain, thalamus) and limbic (hippocampus, amygdala, anterior cingulate cortex) regions. Results: No significant group differences were observed in [11C]McN 5652 BP or V3⬙ in the ROIs. No significant group differences were detected in the voxelwise analysis of BP or V3⬙ maps. Conclusions: OCD without comorbid depression, may not be associated with major changes in SERT availability in subcortical and limbic regions. Biol Psychiatry 2003;54: 1414 –1421 © 2003 Society of Biological Psychiatry Key Words: Obsessive-compulsive disorder, serotonin transporters, [11C]McN 5652, positron emission tomography, caudate nucleus From the Departments of Psychiatry (HBS, IL, MS, HYH, D-RH, AA-D, MRL, ML) and Radiology (HYH, D-RH, ML), Columbia University, College of Physicians and Surgeons, and from the Anxiety Disorders Clinic (HBS, MRL) and Division of Functional Brain Mapping (IL, MS, HYH, D-RH, AA-D, ML), New York State Psychiatric Institute, New York, New York. Address reprint requests to Dr. Simpson, Anxiety Disorders Clinic, Unit 69, 1051 Riverside Drive, New York NY 10032. Received January 27, 2003; revised April 22, 2003; accepted May 12, 2003.
© 2003 Society of Biological Psychiatry
Introduction
O
bsessive– compulsive disorder (OCD) is characterized by intrusive and distressing thoughts, images, or impulses (i.e., obsessions) and by repetitive mental or behavioral acts (i.e., compulsions). Findings from functional brain imaging studies have led to the hypothesis that obsessions and compulsions result from hyperactivity in a brain circuit involving the orbitofrontal cortex, caudate nucleus, and thalamus (Saxena et al 2001). Dysfunction in serotonin (5-HT) neurotransmission has also been implicated in OCD (Barr et al 1992; Rauch et al 2002), based on the selective efficacy of 5-HT reuptake inhibitors (SRIs) in this illness, abnormalities of some peripheral measures of 5-HT function (i.e., reduced density of platelet 5-HT transporters) in OCD, and behavioral and physiologic responses in OCD subjects following challenges with certain 5-HT probes (i.e., meta-chlorophenylpiperazine). Because these are indirect measures of brain function, however, it remains unknown whether OCD involves a primary defect in the brain 5-HT system or whether SRIs reduce OCD severity because 5-HT brain pathways modulate dysfunction in other brain circuits. In vivo imaging of 5-HT transporters (SERT) with positron emission tomography (PET) offers a unique tool to probe 5-HT function. Not only is SERT a critical protein for the regulation of 5-HT and the initial site of SRI action, but because of SERT’s localization on 5-HT nerve terminals, SERT density has been used as a marker for the number or integrity of 5-HT nerve terminals. Trans-1,2,3,5,6,10 -hexahydro-6-[4-(methylthio)phenyl] pyrrolo-[2,1-a]-isoquinoline (McN 5652) is a potent (KD ⫽ .4 nmol/L) inhibitor of 5-HT uptake (Shank et al 1988). Labeled with C-11, McN5652 has been successfully developed as a PET radiotracer for SERT imaging (Suehiro et al 1993). The brain uptake of [11C]McN 5652 is consistent with the known distribution of SERT, and its specific binding is selectively blocked by pretreatment with SERT inhibitors (Parsey et al 2000; Szabo et al 1995). Other researchers have used [11C]McN 5652 to 0006-3223/03/$30.00 doi:10.1016/S0006-3223(03)00544-4
Serotonin Transporters in OCD
study SERT brain distribution in ecstasy abusers (McCann et al 1998) and patients with major depression (Ichimiya et al 2002). Because of its high nonspecific binding, [11C]McN 5652 cannot reliably measure SERT in areas of low SERT density (i.e., the cortex); however, it can measure SERT in subcortical and limbic areas (Parsey et al 2000). This study compared subcortical and limbic brain SERT availability in adult OCD and matched healthy control subjects using PET and [11C]McN 5652. We hypothesized that OCD subjects would have decreased SERT availability compared with healthy control subjects. Decreased SERT availability was hypothesized because platelet studies have found reduced SERT density in OCD and because some have proposed that OCD is characterized by decreased 5-HT neurotransmission (which could result from decreased 5-HT nerve terminals and thus be associated with decreased SERT density). The caudate nucleus was our initial focus because some structural and functional brain imaging studies have found caudate abnormalities in OCD subjects leading to hypotheses of caudate dysfunction in OCD (Saxena et al 2001). However, because this was the first PET SERT study in OCD, SERT availability was also measured in other subcortical and limbic regions, and the region of interest analysis was complemented by an exploratory voxelwise analysis on SERT parametric maps of the whole brain.
BIOL PSYCHIATRY 2003;54:1414 –1421
1415
Radiochemistry In this article, [11C]McN 5652 was prepared as previously described (Parsey et al 2000) and is used to designate the active enantiomer [11C](⫹)McN 5652.
PET Procedures The [11C]McN 5652 PET scanning was performed with the ECAT EXACT HR⫹ (Siemens/CTI, Knoxville, Tennessee) operated in the three-dimensional (3-D) mode. The plane and axial resolutions were 4.4 and 4.1 mm full width half-maximum (FWHM) at the center of the field of view. Following a 10-min transmission scan, [11C]McN 5652 was injected intravenously over 45 sec. Emission data were collected for 130 min, as 22 frames of increasing duration. Previous analysis of [11C]McN 5652 regional uptake established that 130 min of data were required for stable estimates of SERT availability (Parsey et al 2000). Images were reconstructed to a 128 ⫻ 128 matrix (pixel size of 2.5 ⫻ 2.5 mm2), attenuation corrected and filtered with a Sheppe .5 filter. After reconstruction, the final PET image resolution was 5.1 mm FWHM at the center of the field of view.
Input Function Measurement Following [11C]McN 5652 injection, arterial samples were collected every 5 sec with an automated sampling system for the first 2 min and manually thereafter at longer intervals (total ⫽ 32 samples). Metabolite correction was performed with high-pressure liquid chromatography (Parsey et al 2000). Determination of [11C]McN 5652 free fraction in the plasma (f1) with ultracentrifugation technique is impaired by very high [11C]McN 5652 filter retention (Parsey et al 2000) and therefore was not obtained.
Methods and Materials Sample Selection
Magnetic Resonance Images
The institutional review boards of the New York State Psychiatric Institute and Columbia University approved this study. Written informed consent was obtained from each subject after explanation of study procedures. Subjects were recruited by media advertisements and word of mouth. Participants were between the ages of 18 and 55, had no significant medical problems, were not pregnant or nursing, had no current or past neurologic disorder, and were free of psychoactive medications for at least 3 weeks (6 weeks for fluoxetine). The OCD subjects fulfilled OCD criteria from DSM-IV for at least 1 year and had no other current Axis I psychiatric disorder (other than social phobia in one subject). Healthy control subjects had no current or past DSM-IV Axis I psychiatric disorder. Groups were matched for age, gender, and ethnicity. Diagnoses were made by clinical interview and confirmed with the Structured Clinical Interview for DSM-IV. Medical health was confirmed by physical exam, routine blood tests, urinalysis, urine drug screen, and electrocardiogram. On the day of the PET scan, OCD and depressive severity were assessed by a trained rater using the Yale–Brown Obsessive–Compulsive Scale (Y-BOCS, scale range: 0 – 40; Goodman et al 1989a, 1989b) and the Hamilton Depression Rating Scale (HAM-D, 17-item, scale range: 0 –50; Hamilton 1960).
The MRIs (3-D spoiled gradient recalled acquisition in the steady state) were acquired on a GE 1.5-T Signa Advantage system and processed as previously described (Abi-Dargham et al 2002).
Model and Outcome Measures Brain activity was corrected for plasma activity assuming a 5% blood volume. Derivation of [11C]McN 5652 distribution volumes was achieved by kinetic analysis, using the metabolitecorrected arterial time activity curve as input function. Kinetic analysis was based on a one tissue compartment model, shown by several groups to be the optimal model for [11C]McN 5652 (Buck et al 2000; Parsey et al 2000; Szabo et al 1999). This model includes the arterial plasma compartment and one tissue compartment. The latter is the sum of the nondisplaceable (i.e., intracerebral free and nonspecifically bound radiotracer) and the specific binding (i.e., specifically bound tracer) compartments. In this model, tissue tracer concentration, CT(t), is related to plasma tracer concentration, CA(t), by
dC T (t)/dt ⫽ K 1C A(t) ⫺ k 2C T(t) ⫺1
⫺1
(1)
where K1 (mL g min ) is the fractional rate constant for radiotracer delivery from the arterial plasma compartment to the
1416
H.B. Simpson et al
BIOL PSYCHIATRY 2003;54:1414 –1421
brain compartment, and k2 (min⫺1) is the fractional rate constant for the efflux from the brain to the arterial plasma. The analytical solution of Equation 1 is given by the convolution (R) of the arterial input function with one exponential function, with K1 as intercept and k2 as exponent:
Kinetic analysis was performed with iterative nonlinear regression using a Levenberg-Marquart least squares minimization procedure implemented in MATLAB (The Math Works, South Natick, Massachusetts). The minimization procedure was weighted by the frame acquisition time.
C T (t) ⫽ K 1e ⫺k2t R C A(t)
VOXELWISE ANALYSIS. The MRI coregistered PET data were resampled to the original PET resolution yielding approximately 2 ⫻ 105 voxels per brain volume. Kinetic analysis was performed on each voxel using the same one tissue compartment model as in the ROI analysis and a basis function method (Gunn et al 1997). A set of exponential functions of the form exp(⫺k2t) were convolved with the arterial plasma input function to create a set of 50 basis functions. The k2 values were uniformly logarithmically spaced and drawn from a physiologically plausible range determined from the ROI based analysis. At each voxel, a value of K1 was determined for each basis function (i.e., for each value of k2) as the K1 value that minimized the weighted squared difference between model and data. Of the set of 50, the basis function with the smallest residual was selected, and VT was computed as K1/ k2. BP and V3⬙ parametric maps were then created for each subject by applying Equations 4 and 5 to each voxel, where VT(CER) was the mean of all the cerebellar voxelwise estimates for that subject.
(2)
Total distribution volume (VT, mL of plasma gm⫺1 of tissue) was obtained as
V T ⫽ K 1/k 2
(3)
Two measures of SERT availability were calculated. First, the binding potential (BP, mL g⫺1) was derived as the difference between VT in the ROI or voxel of interest (VOI) and VT in the cerebellum, given the low to negligible density of SERT in this region (Ba¨ ckstro¨ m et al 1989; Cortes et al 1988; Laruelle et al 1988):
BP ⫽ V T (ROI or VOI) ⫺ V T(CER)
(4)
Second, the equilibrium specific to nonspecific partition coefficient (V3⬙, unitless) was derived as the ratio of BP to cerebellar VT:
V 3⬙ ⫽ BP/V T(CER)
(5)
The BP and V3⬙ are related to transporter parameters as follows (Laruelle et al 1994):
BP ⫽ f 1B max/K D
(6)
V 3⬙ ⫽ f 2B max/K D ⫽ f 1B max/K DV T(CER)
(7)
where Bmax is the regional concentration of SERT (pmoles g⫺1 of tissue), KD is the affinity of [11C]McN 5652 for SERT (nmol/L), f1is the free fraction in the plasma, and f2 is the free fraction in the nondisplaceable compartment.
Image Analysis PREPROCESSING. Image analysis was performed blind to diagnosis using MEDx (Sensor Systems, Sterling, Virginia). To correct for head movement, PET frames were realigned to a frame of reference using automated image registration (Woods et al 1992) and coregistered to the MRI. ROI ANALYSIS. The ROI boundaries and one region of reference (cerebellum) were drawn on each subject’s MRI; ROIs included dorsal caudate, dorsal putamen, ventral striatum, midbrain, thalamus, hippocampus, amygdala, and anterior cingulate cortex. The midbrain includes various structures with dense SERT concentration (raphe nuclei, substantia nigra, locus ceruleus, ventral tegmental area, and superior and inferior colliculi). Mawlawi et al (2001) describe and illustrate the criteria for delineating the boundaries of striatal substructures (dorsal caudate, dorsal putamen, and ventral striatum, the latter including nucleus accumbens, ventral caudate, and ventral putamen) as well the ability of quantifying radioactive signal from these substructures. Criteria for delineating the other ROIs are available upon request. Right and left ROI volumes were summed.
Statistical Analysis ROI ANALYSIS. Volume, K1, VT, BP, and V3⬙ were each analyzed with univariate repeated-measures analysis of variance (ANOVA); brain region (ROIs and cerebellum) was the withinsubjects factor, and diagnostic group was the between-subjects factor. This model was used because multiple brain regions were examined in the same subject, and it can test for correlations between repeated measures. When the tests of sphericity were significant, the Huynh–Feldt correction for the degrees of freedom was used. Exploratory analyses also compared OCD subjects and control subjects on each parameter within each brain region using two sample unpaired t tests. Within the OCD group, the relationship between SERT availability and OCD severity (measured by the Y-BOCS) was examined with the Pearson’s Product–Moment correlation. All statistical tests were two-tailed with level of significance ␣ ⫽ .05. VOXELWISE ANALYSIS. All voxelwise analyses were performed in the Statistical Parametric Mapping 99 (SPM, Wellcome Department of Cognitive Neurology, University College, London, United Kingdom) environment. The MRIs were normalized to the Montreal Neurologic Institute T1 template. The same transformation was applied to the parametric maps, yielding parametric data normalized into a common template space. These data were smoothed with a Gaussian kernel (FWHM ⫽ 6 mm) to comply with the random fields smoothness assumption employed by the SPM multiple comparisons correction procedure. Two sample unpaired t tests were used to test for group differences in K1, VT, BP, and V3⬙ maps (voxel and cluster significance level p ⬍ .05, corrected for multiple comparisons). Within the OCD group, the linear correlation between OCD severity (measured by the Y-BOCS) and voxelwise BP and V3⬙ was examined.
Serotonin Transporters in OCD
Table 2. Regional Volumes (mm3)
Table 1. Demographic and Clinical Characteristics Parameter Number Mean Age in Years (SD) Number of Men/Women Ethnicity History of SRI Treatment Response to SRI Mean weeks (SD) since last SRI dose Mean Age of OCD Onset in Years (SD) Mean Y-BOCS (SD) Mean HAM-D (SD) DSM-IV Axis I Psychiatric Comorbidity No current or past History of MDD only Subthreshold trichotillomania and history of MDD Both MDD and PTSD Social phobia and history of anorexia History of cannabis abuse History of specific phobia Current/History of Isolated Tics
1417
BIOL PSYCHIATRY 2003;54:1414 –1421
Comparison Subjects
OCD Patients
11 11 31 (10) 31 (12) 6/5 6/5 4AA/5Cau/2H 5AA/5Cau/1H — 6 of 11 — 5 of 6 — 65 (73) — 17 (6) — 20 (4) — 6 (4) 11 — —
3 3 1
— — — — —
1 1 1 1 2/1
AA, African American; Cau, Caucasian; HAM-D, Hamilton Depression Rating Scale; H, Hispanic; SRI, serotonin reuptake inhibitor; MDD, major depressive disorder; OCD, obsessive-compulsive disorder; PTSD, posttraumatic stress disorder; Y-BOCS, Yale–Brown Obsessive–Compulsive Scale.
Results Sample Characteristics More than 120 OCD subjects were screened. Most were ineligible because they were on psychiatric medication or had an excluded current comorbid psychiatric condition (e.g., major depressive disorder (MDD), other excluded anxiety disorder, substance or alcohol abuse). Of the 34 eligible OCD subjects, 23 declined participation, primarily because of concerns about the arterial line and radiation exposure. Demographic and clinical characteristics of the 11 comparison and 11 OCD subjects who participated are shown in Table 1. There were no significant group differences in mean age (t ⫽ .00, df ⫽ 20, p ⫽ 1.00), gender distribution (Fisher’s Exact Test, p ⫽ 1.00), or ethnicity (Fisher’s Exact Test, p ⫽ 1.00). As shown in Table 1, OCD subjects had on average moderate OCD severity: one had mild OCD (Y-BOCS ⫽ 12); most had moderate OCD severity (Y-BOCS ⫽ 16 –23); two had more severe OCD (Y-BOCS ⫽ 24 –29); and none had extreme OCD (Y-BOCS ⱖ32). Although one subject had predominantly symmetry and exactness obsessions with ordering or arranging compulsions, the other 10 endorsed at least two categories of obsessions and two categories of compulsions on the Y-BOCS Symptom Checklist. Within the five-factor model of symptom subtypes (Mataix-Cols et al 1999), these 10 subjects endorsed
Region Cerebellum Dorsal Caudate Dorsal Putamen Ventral Striatum Midbrain Thalamus Hippocampus Amygdala Anterior Cingulate
Comparison Subjects
OCD Subjects
t
p
64,813 (7890) 5519 (595) 8430 (1260) 2042 (654) 6396 (616) 8391 (1698) 7416 (983) 3598 (586) 7101 (1767)
62,750 (13,233) 5300 (1093) 9343 (1865) 2688 (1125) 6782 (844) 8384 (1144) 8019 (1418) 4590 (1188) 7203 (1673)
.44 .58 1.35 1.65 1.23 .01 1.16 2.48 .14
.66 .57 .19 .11 .23 .99 .26 .02 .89
Values are mean (SD), n ⫽ 11 and 11 for comparison and OCD subjects. t and p are value and probability of unpaired two sample t tests (df ⫽ 20), respectively. Right and left volumes were summed. OCD, obsessive– compulsive disorder.
as their current major problem different combinations of the following four factors: contamination obsessions and washing compulsions (n ⫽ 7); aggressive obsessions and checking compulsions (n ⫽ 5); symmetry and exactness obsessions and ordering or arranging compulsions (n ⫽ 7); hoarding obsessions and compulsions (n ⫽ 7). Seven also had prominent mental rituals. Five OCD subjects were SRI naive; the other six had been off SRI medication for an average of 65 weeks (SD ⫽ 73; range: 6 –188 weeks). Three reported infrequent involuntary motor movements that were consistent with isolated motor tics (current, n ⫽ 2; past, n ⫽ 1). None met current or past DSM-IV criteria for a tic disorder.
Injected Dose and Input Function There were no significant group differences in mean injected dose (control subjects, 11.28 ⫾ 4.86 mCi; OCD subjects, 13.51 ⫾ 3.50 mCi; t ⫽ 1.23, df ⫽ 20, p ⫽ .23), mean specific activity at time of injection (control subjects, 880 ⫾ 695 Ci/mmoles; OCD subjects, 834 ⫾ 349 Ci/mmol; t ⫽ .20, df ⫽ 20, p ⫽ .72), injected mass (control subjects, 4.59 ⫾ 1.51 g; OCD subjects, 5.07 ⫾ .91 g; t ⫽ .89, df ⫽ 20, p ⫽ .38) or rate of plasma clearance (control subjects, 147 ⫾ 39 L h⫺1; OCD subjects, 113 ⫾ 52 L h⫺1; t ⫽ 1.72, df ⫽ 20, p ⫽ .10).
Regional Volumes Volumes of ROIs are presented in Table 2. Repeatedmeasures ANOVA showed a significant effect of region (F ⫽ 557, df ⫽ 1.23, 24.55, p ⬍ .001), no effect of group (F ⫽ .083, df ⫽ 1, 20, p ⫽ .78), and no group-by-region interaction (F ⫽ .32, df ⫽ 1.23, 24.55, p ⫽ .62). Regionwise comparisons showed no statistically significant group differences in ROI volumes, with one exception: OCD subjects had significantly larger amygdala than control subjects, a difference that did not survive Bonferroni correction (n ⫽ 9 regions).
1418
H.B. Simpson et al
BIOL PSYCHIATRY 2003;54:1414 –1421
Table 3. Regional Delivery (K1, mL g⫺1 min⫺1) Region Cerebellum Dorsal Caudate Dorsal Putamen Ventral Striatum Midbrain Thalamus Hippocampus Amygdala Anterior Cingulate
Comparison Subjects
OCD Subjects
t
p
.30 (.04) .32 (.04) .36 (.05) .31 (.05) .26 (.03) .33 (.05) .24 (.04) .22 (.03) .32 (.05)
.33 (.08) .34 (.07) .39 (.08) .35 (.08) .27 (.06) .35 (.07) .24 (.04) .26 (.05) .36 (.08)
1.14 .95 1.22 1.34 .51 .69 .70 .98 1.33
.27 .36 .24 .20 .61 .50 .49 .34 .20
Values are mean (SD), n ⫽ 11 and 11 for comparison and OCD subjects. t and p are value and probability of two sample t tests (df ⫽ 20), respectively. OCD, obsessive– compulsive disorder.
ROI Kinetic Analysis Kinetic analysis of time activity curves converged in all cases. K1 values are listed in Table 3. Repeated-measures ANOVA showed a significant effect of region (F ⫽ 140, df ⫽ 5.14, 102.86, p ⬍ .001), no effect of group (F ⫽ 1.11, df ⫽ 1, 20, p ⫽ .31), and no significant group-by-region interaction (F ⫽ 1.79, df ⫽ 5.14, 102.86, p ⫽ .12). Regionwise comparisons revealed no significant group differences in K1 values in any ROI. The VT values are presented in Figure 1. Repeatedmeasures ANOVA showed a significant effect of region (F ⫽ 90, df ⫽ 2.27, 45.35, p ⬍ .001), no effect of group (F ⫽ .04, df ⫽ 1, 20, p ⫽ .85), and no significant group-by-region interaction (F ⫽ .11, df ⫽ 2.27, 45.35, p ⫽ .92). No group differences were observed in cerebellar VT (VT(CER); control subjects: 18.66 ⫾ 3.06 mL g⫺1; OCD subjects: 19.26 ⫾ 3.97 mL g⫺1, t ⫽ .39, df ⫽ 20, p ⫽ .70).
Regional SERT Availability The BP values are presented in Table 4. Repeatedmeasures ANOVA showed a significant effect of region (F ⫽ 73, df ⫽ 2.56, 51.27, p ⬍ .001), no effect of group (F ⫽ .003, df ⫽ 1, 20, p ⫽ .96), and no significant group-by-region interaction (F ⫽ .13, df ⫽ 2.56, 51.27, p ⫽ .92). Regionwise comparisons revealed no significant group differences in BP values in any ROI. The V3⬙ values are presented in Table 5. Repeatedmeasures ANOVA showed a significant effect of region (F ⫽ 108, df ⫽ 3.52, 70.59, p ⬍ .001), no effect of group (F ⫽ .29, df ⫽ 1, 20, p ⫽ .60), and no significant group-by-region interaction (F ⫽ .30, df ⫽ 3.53, 70.59, p ⫽ .86). No significant between-group differences in V3⬙ values were found in the ROIs. There were no significant correlations between OCD severity and [11C]McN 5652 BP or V3⬙ in any ROI. Moreover, post-hoc analyses found no significant mean differences (Mann–Whitney U, all p values ⬎ .25) in ROI
Figure 1. Mean [11C]McN 5652 total distribution volumes (VT, mL g⫺1) and standard deviations in comparison subjects (n ⫽ 11) and subjects with obsessive– compulsive disorder (OCD) (n ⫽ 11) in cerebellum (CER) and regions of interest. Repeatedmeasures analysis of variance showed a significant effect of region (F ⫽ 90, df ⫽ 2.27, 45.35, p ⬍ .001), no effect of group (F ⫽ .04, df ⫽ 1, 20, p ⫽ .85), and no significant group-byregion interaction (F ⫽ .11, df ⫽ 2.27, 45.35, p ⫽ .92). DCA, dorsal caudate; DPU, dorsal putamen; VST, ventral striatum; MID, midbrain; THA, thalamus; AMY, amygdala; HIP, hippocampus; CIN, anterior cingulate.
BP or V3⬙ in OCD patients who were SRI naive (n ⫽ 5) or not (n ⫽ 6) or between OCD patients with a history of tics (n ⫽ 3) and those without (n ⫽ 8).
Voxelwise Analysis Mean K1 maps are shown in Figure 2. In the SPM analysis of the K1 maps, no voxels nor clusters of voxels were Table 4. Regional Binding Potentials (BP, mL g⫺1) Region
Comparison Subjects
OCD Subjects
t
p
Dorsal Caudate Dorsal Putamen Ventral Striatum Midbrain Thalamus Hippocampus Amygdala Anterior Cingulate
18.03 (6.60) 21.78 (6.28) 22.78 (5.96) 29.45 (8.08) 17.37 (4.50) 6.90 (2.23) 20.00 (7.73) 4.23 (1.78)
17.06 (9.16) 23.36 (10.84) 22.64 (10.10) 29.70 (15.19) 17.45 (6.88) 6.97 (3.32) 19.91 (11.00) 4.60 (2.15)
.29 .42 .04 .05 .03 .06 .02 .44
.78 .68 .97 .96 .97 .95 .98 .66
Values are mean (SD), n ⫽ 11 and 11 for comparison and OCD subjects. t and p are value and probability of unpaired two sample t tests (df ⫽ 20), respectively. OCD, obsessive– compulsive disorder.
Serotonin Transporters in OCD
BIOL PSYCHIATRY 2003;54:1414 –1421
1419
Table 5. Equilibrium Specific to Nonspecific Partition Coefficient (V3⬙, unitless) Region Dorsal Caudate Dorsal Putamen Ventral Striatum Midbrain Thalamus Hippocampus Amygdala Anterior Cingulate
Comparison Subjects
OCD Subjects
t
p
.95 (.24) 1.15 (.22) 1.23 (.29) 1.57 (.34) .92 (.17) .37 (.09) 1.04 (.25) .23 (.08)
.84 (.28) 1.17 (.32) 1.14 (.32) 1.49 (.52) .88 (.21) .35 (.13) .99 (.37) .23 (.08)
.90 .12 .70 .44 .54 .27 .43 .23
.38 .90 .49 .67 .60 .79 .67 .82
Values are mean (SD), n ⫽ 11 and 11 for comparison and OCD subjects. t and p are value and probability of unpaired two sample t tests (df ⫽ 20), respectively. OCD, obsessive-compulsive disorder.
significantly different (at p ⬍ .05, corrected for multiple comparisons) between the groups. The SPM analysis of VT maps was negative as well. Mean BP maps are shown in Figure 3. High SERT availability was visualized in midbrain, thalamus, and striatum, in agreement with the ROI analysis and the known distribution of SERT in the human brain (Ba¨ ckstro¨ m et al 1989; Cortes et al 1988; Laruelle et al 1988). In the SPM analysis of the BP maps, no voxels or clusters of voxels were significantly different (at p ⬍ .05, corrected) between the groups. Post hoc analyses of the striatum
Figure 3. Mean parametric maps of binding potential (BP, mL g⫺1) in comparison subjects (n ⫽ 11, top row) and subjects with obsessive– compulsive (OCD) (n ⫽ 11, middle row). Individual BP maps were created as the difference between voxelwise VT and cerebellum VT (i.e., the mean of all the cerebellar voxelwise VT estimates for that subject). Individual BP maps were then spatially normalized to the Montreal Neurologic Institute template (bottom row), and mean BP values for each group were calculated. The BP maps reveal brain areas associated with high serotonin transporter density (i.e., midbrain, thalamus, striatum). No voxels or clusters of voxels were significantly different between the groups in the Statistical Parametric Mapping 99 analysis of these BP maps.
alone (encompassing the three striatal ROIs: dorsal caudate, dorsal putamen, and ventral striatum) demonstrated no voxels nor clusters of voxels significantly different (at p ⬍ .01, uncorrected) between the two groups. Mean V3⬙ maps showed the same pattern of SERT distribution as the BP maps, and the SPM analysis of V3⬙ maps was negative as well. There were no significant correlations between OCD severity and voxelwise BP or V3⬙.
Discussion
Figure 2. Mean parametric maps of tracer delivery to the brain (K1, mL g⫺1 min⫺1) in comparison subjects (n ⫽ 11, top row) and subjects with obsessive-compulsive disorder (OCD) (n ⫽ 11, middle row). Individual K1 maps were derived and spatially normalized to the Montreal Neurologic Institute template (bottom row), and mean K1 values for each group were calculated. No voxels or clusters of voxels were significantly different between the groups in the Statistical Parametric Mapping 99 analysis of these K1 maps. OCD, obsessive-compulsive disorder.
This study compared SERT availability in the brains of subjects with OCD and healthy subjects. Contrary to the hypothesis, there were no significant group differences in SERT availability in the caudate nucleus or any of the ROIs. There also were no significant group differences in SERT binding in the SPM analyses. Within the OCD group, OCD severity and SERT availability were not significantly correlated in either the ROI or the voxelwise analyses. This study had several strengths. First, measurement of the arterial input function enabled the quantitative derivation of distribution volumes, from which both BP and V3⬙
1420
H.B. Simpson et al
BIOL PSYCHIATRY 2003;54:1414 –1421
were calculated. Although methods for deriving [11C]McN 5652 V3⬙ without arterial sampling have been proposed (Ichimiya et al 2002), only with the arterial input function could we demonstrate the absence of group differences in cerebellum distribution volumes and thus validate the use of V3⬙ for between-group comparisons of receptor parameters. Second, ROI and voxel-based analyses were conducted. The ROI method has the advantage of being hypothesis driven, but its ability to detect between-group differences in parameters of interest is limited to differences occurring in anatomically predefined regions. In contrast, the voxelwise approach enables the detection of differences in spatial clusters that include only a fraction of an ROI or that overlap different ROIs. In this study, the absence of SERT alterations in the ROIs of OCD subjects was complemented by an equally negative voxelwise analysis. Third, the OCD sample studied was relatively uncontaminated by SRI treatment: five were SRI naive, and the other six had been free of SRIs for an average of 65 weeks. The study also had several limitations. First, the sample was relatively small (11 OCD and 11 comparisons subjects). The between-subject variability (coefficient of variation) in SERT V3⬙ in the eight ROIs was 24 ⫾ 6% in comparison subjects and 30 ⫾ 7% in OCD subjects, providing adequate power (⬎.80) to detect only relatively large differences in SERT availability. For example, in the dorsal caudate, a decrease of 31% or more in V3⬙ in OCD subjects was required to detect a significant group difference. Nonetheless, there was no trend toward betweengroup differences in the entire sample. The largest difference observed, an 11% decrease in dorsal caudate V3⬙ in OCD subjects, had an effect size of –.38; more than 85 subjects in each group would be needed for this difference to reach significance at the p ⬍ .05 level. On the other hand, we note that one OCD subject, who was not an outlier with respect to the entire sample, was an outlier with respect to the OCD sample for some outcome measures in some ROIs (i.e., BP or V3⬙ ⬎ [upper quartile ⫹ 1.5 ⫻ (interquartile range)]). In post-hoc analyses that excluded this subject, the largest observed difference was then an 18% decrease in dorsal caudate V3⬙ in OCD subjects (t ⫽ 1.72, df ⫽ 19, p ⫽ .10) with an effect size of –.72; the p value still exceeded .25 for all other group comparisons of BP and V3⬙. There was no justification to exclude this patient a priori from the analyses based on experimental data, and there were no obvious clinical differences between this subject and the other OCD subjects. Given the clinical heterogeneity of OCD, these negative results do not exclude the possibility of SERT alterations in subtypes of OCD not included in our sample. For example, because of the need for a drug-free interval before the study (coupled with fears about an arterial line and radiation exposure), no subjects with extreme OCD
severity could be enrolled. Moreover, only one subject reported onset of OCD in childhood, none were known to be SRI unresponsive, none had a chronic tic disorder, and most had a range of OCD symptoms including mental compulsions, as is typical of OCD subjects (Foa and Kozak 1995). Thus, a second limitation of this study is that the findings do not rule out alterations in SERT availability confined to specific clinical profiles not included in this study (e.g., extreme OCD, childhood-onset OCD, SRIunresponsive OCD, comorbid chronic tic disorder, pure hoarders). On the other hand, our findings suggest that OCD in general is not characterized by major decreases in SERT availability in the examined brain regions. Future brain SERT studies might consider studying particular subtypes of OCD. Third, because [11C]McN 5652 is affected by high nonspecific binding, SERT measurement in regions of low SERT density, such as the human neocortex, is unreliable with this tracer. Therefore, the ROI analysis was restricted to regions in which the derivation of SERT BP and V3⬙ has been found reliable (Parsey et al 2000). Thus, this study does not exclude the existence of SERT alterations in OCD in the orbitofrontal (or other parts of the) cortex. The recent development of superior SERT PET radiotracers such as [11C]DASB (Wilson et al 2000) and [11C]AFM (Huang et al 2001, 2002) might permit investigations of the neocortex in the future. An additional benefit of these new tracers is that their free fraction can be measured, allowing one to control for this parameter. In conclusion, this study failed to detect marked differences in SERT availability in subcortical and limbic regions in untreated OCD subjects. Further evaluation of SERT availability in the neocortex and in specific OCD subtypes is desirable to complete this assessment. These results do not exclude the existence of altered 5-HT transmission in OCD that is not associated with SERT density alterations, such as abnormalities of 5-HT release or of 5-HT receptors. Thus, investigations of other aspects of the serotonergic system in OCD subjects are warranted.
This work was supported by the Obsessive-Compulsive Foundation and the National Institute of Mental Health (Grant No. K23 MH01907 to HBS). We thank Gordon Frankle, Wendy Garcia, Ingrid Gelbard-Stokes, Elizabeth Hackett, Leyla Khenissi, Shu-Hsing Lin, Kimchung Ngo, Julie Montoya, Chaka Peters, Norm Simpson, and Kris Wolff for excellent technical assistance.
References Abi-Dargham A, Mawlawi O, Lombardo I, Gil R, Martinez D, Huang Y, et al (2002): Prefrontal dopamine D1 receptors and working memory in schizophrenia. J Neurosci 22:3708 – 3719.
Serotonin Transporters in OCD
Ba¨ ckstro¨ m I, Bergstro¨ m M, Marcusson J (1989): High affinity [3H]paroxetine binding to serotonin uptake sites in human brain tissue. Brain Res 486:261–268. Barr LC, Goodman WK, Price LH, McDougle CJ, Charney DS (1992): The serotonin hypothesis of obsessive compulsive disorder: Implications of pharmacologic challenge studies. J Clin Psychiatry 53(suppl):17–28. Buck A, Gucker PM, Schonbachler RD, Arigoni M, Kneifel S, Vollenweider FX, et al (2000): Evaluation of serotonergic transporters using PET and [11C](⫹)McN-5652: Assessment of methods. J Cereb Blood Flow Metab 20:253–262. Cortes R, Soriano E, Pazos A, Probst A, Palacios JM (1988): Autoradiography of antidepressant binding sites in the human brain: Localization using [3H] [3H]imipramine and [3H] [3H]paroxetine. Neuroscience 27:473–496. Foa EB, Kozak MJ (1995): DSM-IV field trial: Obsessivecompulsive disorder. Am J Psychiatry 152:90 –96. Goodman WK, Price LH, Rasmussen SA, Mazure C, Delgado P, Heninger GR, et al (1989b): The Yale–Brown Obsessive Compulsive Scale. II. Validity. Arch Gen Psychiatry 46:1012–1016. Goodman WK, Price LH, Rasmussen SA, Mazure C, Fleischmann RL, Hill CL, et al (1989a): The Yale–Brown Obsessive Compulsive Scale. I. Development, use, and reliability. Arch Gen Psychiatry 46:1006 –1011. Gunn RN, Lammertsma AA, Hume SP, Cunningham VJ (1997): Parametric imaging of ligand-receptor binding in PET using a simplified reference region model. Neuroimage 6:279 –2787. Hamilton M (1960): A rating scale for depression. J Neurol Neurosurg Psychiatry 23:56 – 62. Huang Y, Bae SA, Zhu Z, Guo N, Hwang DR, Laruelle M (2001): Fluorinated analogues of ADAM as new PET radioligands for the serotonin transporter: Synthesis and pharmacological evaluation. J Labelled Compd Radiopharm 44:S18 –S20. Huang Y, Hwang DR, Narendran R, Sudo Y, Chatterjee R, Bae SA, et al (2002): Comparative evaluation in nonhuman primates of five PET radiotracers for imaging the serotonin transporters: [11C]McN 5652, [11C]ADAM, [11C]DASB, [11C]DAPA, and [11C]AFM. J Cereb Blood Flow Metab 22:1377–1398. Ichimiya T, Suhara T, Sudo Y, Okubo Y, Nakayama K, Nankai M, et al (2002): Serotonin transporter binding in patients with mood disorders: A PET study with [11C](⫹)McN5652. Biol Psychiatry 51:715–722. Laruelle M, van Dyck C, Abi-Dargham A, Zea-Ponce Y, Zoghbi SS, Charney DS, et al (1994): Compartmental modeling of iodine-123-iodobenzofuran binding to dopamine D2 receptors in healthy subjects. J Nucl Med 35:743–754. Laruelle M, Vanisberg M, Maloteaux J (1988): Regional and subcellular localization in human brain of [3H]paroxetine
BIOL PSYCHIATRY 2003;54:1414 –1421
1421
binding, a marker of serotonin uptake sites. Biol Psychiatry 24:299 –309. Mataix-Cols D, Rauch SL, Manzo PA, Jenike MA, Baer L (1999): Use of factor-analyzed symptom dimensions to predict outcome with serotonin reuptake inhibitors and placebo in the treatment of obsessive-compulsive disorder. Am J Psychiatry 156:1409 –1416. Mawlawi O, Martinez D, Slifstein M, Broft A, Chatterjee R, Hwang DR, et al (2001): Imaging human mesolimbic dopamine transmission with positron emission tomography: I. Accuracy and precision of D2 receptor parameter measurements in ventral striatum. J Cereb Blood Flow Metab 21:1034 –1057. McCann UD, Szabo Z, Scheffel U, Dannals RF, Ricaurte GA (1998): Positron emission tomographic evidence of toxic effect of MDMA (“Ecstasy”) on brain serotonin neurons in human beings. Lancet 352:1433–1437. Parsey RV, Kegeles LS, Hwang DR, Simpson N, Abi-Dargham A, Mawlawi O, et al (2000): In vivo quantification of brain serotonin transporters in humans using [11C]McN 5652. J Nucl Med 41:1465–1477. Rauch SL, Cora-Locatelli G, Greenberg BD (2002): Pathogenesis of obsessive-compulsive disorder. In: Stein DJ, Hollander E, editors. Textbook of Anxiety Disorders. Washington, DC: American Psychiatric Association, 191–205. Saxena S, Bota RG, Brody AL (2001): Brain-behavior relationships in obsessive-compulsive disorder. Semin Clin Neuropsychiatry 6:82–101. Shank RP, Vaught JL, Pelley KA, Setler PE, McComsey DF, Maryanoff BE (1988): McN-5652: A highly potent inhibitor of serotonin uptake. J Pharmacol Exp Ther 247:1032–1038. Suehiro M, Scheffel U, Dannals RF, Ravert HT, Ricaurte GA, Wagner H Jr (1993): A PET radiotracer for studying serotonin uptake sites: Carbon-11-McN-5652Z. J Nucl Med 34:120 –127. Szabo Z, Kao PF, Scheffel U, Suehiro M, Mathews WB, Ravert HT, et al (1995): Positron emission tomography imaging of serotonin transporters in the human brain using [11C](⫹)McN5652. Synapse 20:37–43. Szabo Z, Scheffel U, Mathews WB, Ravert HT, Szabo K, Kraut M, et al (1999): Kinetic analysis of [11C]McN5652: A serotonin transporter radioligand. J Cereb Blood Flow Metab 19:967–981. Wilson AA, Ginovart N, Schmidt M, Meyer JH, Threlkeld PG, Houle S (2000): Novel radiotracers for imaging the serotonin transporter by positron emission tomography: Synthesis, radiosynthesis, and in vitro and ex vivo evaluation of (11)Clabeled 2-(phenylthio)araalkylamines. J Med Chem 43:3103– 3110. Woods RP, Cherry SR, Mazziotta JC (1992): Rapid automated algorithm for aligning and reslicing PET images. J Comp Assist Tomogr 16:620 –633.
© 2003 Society of Biological Psychiatry
Introduction
O
bsessive– compulsive disorder (OCD) is characterized by intrusive and distressing thoughts, images, or impulses (i.e., obsessions) and by repetitive mental or behavioral acts (i.e., compulsions). Findings from functional brain imaging studies have led to the hypothesis that obsessions and compulsions result from hyperactivity in a brain circuit involving the orbitofrontal cortex, caudate nucleus, and thalamus (Saxena et al 2001). Dysfunction in serotonin (5-HT) neurotransmission has also been implicated in OCD (Barr et al 1992; Rauch et al 2002), based on the selective efficacy of 5-HT reuptake inhibitors (SRIs) in this illness, abnormalities of some peripheral measures of 5-HT function (i.e., reduced density of platelet 5-HT transporters) in OCD, and behavioral and physiologic responses in OCD subjects following challenges with certain 5-HT probes (i.e., meta-chlorophenylpiperazine). Because these are indirect measures of brain function, however, it remains unknown whether OCD involves a primary defect in the brain 5-HT system or whether SRIs reduce OCD severity because 5-HT brain pathways modulate dysfunction in other brain circuits. In vivo imaging of 5-HT transporters (SERT) with positron emission tomography (PET) offers a unique tool to probe 5-HT function. Not only is SERT a critical protein for the regulation of 5-HT and the initial site of SRI action, but because of SERT’s localization on 5-HT nerve terminals, SERT density has been used as a marker for the number or integrity of 5-HT nerve terminals. Trans-1,2,3,5,6,10 -hexahydro-6-[4-(methylthio)phenyl] pyrrolo-[2,1-a]-isoquinoline (McN 5652) is a potent (KD ⫽ .4 nmol/L) inhibitor of 5-HT uptake (Shank et al 1988). Labeled with C-11, McN5652 has been successfully developed as a PET radiotracer for SERT imaging (Suehiro et al 1993). The brain uptake of [11C]McN 5652 is consistent with the known distribution of SERT, and its specific binding is selectively blocked by pretreatment with SERT inhibitors (Parsey et al 2000; Szabo et al 1995). Other researchers have used [11C]McN 5652 to 0006-3223/03/$30.00 doi:10.1016/S0006-3223(03)00544-4
Serotonin Transporters in OCD
study SERT brain distribution in ecstasy abusers (McCann et al 1998) and patients with major depression (Ichimiya et al 2002). Because of its high nonspecific binding, [11C]McN 5652 cannot reliably measure SERT in areas of low SERT density (i.e., the cortex); however, it can measure SERT in subcortical and limbic areas (Parsey et al 2000). This study compared subcortical and limbic brain SERT availability in adult OCD and matched healthy control subjects using PET and [11C]McN 5652. We hypothesized that OCD subjects would have decreased SERT availability compared with healthy control subjects. Decreased SERT availability was hypothesized because platelet studies have found reduced SERT density in OCD and because some have proposed that OCD is characterized by decreased 5-HT neurotransmission (which could result from decreased 5-HT nerve terminals and thus be associated with decreased SERT density). The caudate nucleus was our initial focus because some structural and functional brain imaging studies have found caudate abnormalities in OCD subjects leading to hypotheses of caudate dysfunction in OCD (Saxena et al 2001). However, because this was the first PET SERT study in OCD, SERT availability was also measured in other subcortical and limbic regions, and the region of interest analysis was complemented by an exploratory voxelwise analysis on SERT parametric maps of the whole brain.
BIOL PSYCHIATRY 2003;54:1414 –1421
1415
Radiochemistry In this article, [11C]McN 5652 was prepared as previously described (Parsey et al 2000) and is used to designate the active enantiomer [11C](⫹)McN 5652.
PET Procedures The [11C]McN 5652 PET scanning was performed with the ECAT EXACT HR⫹ (Siemens/CTI, Knoxville, Tennessee) operated in the three-dimensional (3-D) mode. The plane and axial resolutions were 4.4 and 4.1 mm full width half-maximum (FWHM) at the center of the field of view. Following a 10-min transmission scan, [11C]McN 5652 was injected intravenously over 45 sec. Emission data were collected for 130 min, as 22 frames of increasing duration. Previous analysis of [11C]McN 5652 regional uptake established that 130 min of data were required for stable estimates of SERT availability (Parsey et al 2000). Images were reconstructed to a 128 ⫻ 128 matrix (pixel size of 2.5 ⫻ 2.5 mm2), attenuation corrected and filtered with a Sheppe .5 filter. After reconstruction, the final PET image resolution was 5.1 mm FWHM at the center of the field of view.
Input Function Measurement Following [11C]McN 5652 injection, arterial samples were collected every 5 sec with an automated sampling system for the first 2 min and manually thereafter at longer intervals (total ⫽ 32 samples). Metabolite correction was performed with high-pressure liquid chromatography (Parsey et al 2000). Determination of [11C]McN 5652 free fraction in the plasma (f1) with ultracentrifugation technique is impaired by very high [11C]McN 5652 filter retention (Parsey et al 2000) and therefore was not obtained.
Methods and Materials Sample Selection
Magnetic Resonance Images
The institutional review boards of the New York State Psychiatric Institute and Columbia University approved this study. Written informed consent was obtained from each subject after explanation of study procedures. Subjects were recruited by media advertisements and word of mouth. Participants were between the ages of 18 and 55, had no significant medical problems, were not pregnant or nursing, had no current or past neurologic disorder, and were free of psychoactive medications for at least 3 weeks (6 weeks for fluoxetine). The OCD subjects fulfilled OCD criteria from DSM-IV for at least 1 year and had no other current Axis I psychiatric disorder (other than social phobia in one subject). Healthy control subjects had no current or past DSM-IV Axis I psychiatric disorder. Groups were matched for age, gender, and ethnicity. Diagnoses were made by clinical interview and confirmed with the Structured Clinical Interview for DSM-IV. Medical health was confirmed by physical exam, routine blood tests, urinalysis, urine drug screen, and electrocardiogram. On the day of the PET scan, OCD and depressive severity were assessed by a trained rater using the Yale–Brown Obsessive–Compulsive Scale (Y-BOCS, scale range: 0 – 40; Goodman et al 1989a, 1989b) and the Hamilton Depression Rating Scale (HAM-D, 17-item, scale range: 0 –50; Hamilton 1960).
The MRIs (3-D spoiled gradient recalled acquisition in the steady state) were acquired on a GE 1.5-T Signa Advantage system and processed as previously described (Abi-Dargham et al 2002).
Model and Outcome Measures Brain activity was corrected for plasma activity assuming a 5% blood volume. Derivation of [11C]McN 5652 distribution volumes was achieved by kinetic analysis, using the metabolitecorrected arterial time activity curve as input function. Kinetic analysis was based on a one tissue compartment model, shown by several groups to be the optimal model for [11C]McN 5652 (Buck et al 2000; Parsey et al 2000; Szabo et al 1999). This model includes the arterial plasma compartment and one tissue compartment. The latter is the sum of the nondisplaceable (i.e., intracerebral free and nonspecifically bound radiotracer) and the specific binding (i.e., specifically bound tracer) compartments. In this model, tissue tracer concentration, CT(t), is related to plasma tracer concentration, CA(t), by
dC T (t)/dt ⫽ K 1C A(t) ⫺ k 2C T(t) ⫺1
⫺1
(1)
where K1 (mL g min ) is the fractional rate constant for radiotracer delivery from the arterial plasma compartment to the
1416
H.B. Simpson et al
BIOL PSYCHIATRY 2003;54:1414 –1421
brain compartment, and k2 (min⫺1) is the fractional rate constant for the efflux from the brain to the arterial plasma. The analytical solution of Equation 1 is given by the convolution (R) of the arterial input function with one exponential function, with K1 as intercept and k2 as exponent:
Kinetic analysis was performed with iterative nonlinear regression using a Levenberg-Marquart least squares minimization procedure implemented in MATLAB (The Math Works, South Natick, Massachusetts). The minimization procedure was weighted by the frame acquisition time.
C T (t) ⫽ K 1e ⫺k2t R C A(t)
VOXELWISE ANALYSIS. The MRI coregistered PET data were resampled to the original PET resolution yielding approximately 2 ⫻ 105 voxels per brain volume. Kinetic analysis was performed on each voxel using the same one tissue compartment model as in the ROI analysis and a basis function method (Gunn et al 1997). A set of exponential functions of the form exp(⫺k2t) were convolved with the arterial plasma input function to create a set of 50 basis functions. The k2 values were uniformly logarithmically spaced and drawn from a physiologically plausible range determined from the ROI based analysis. At each voxel, a value of K1 was determined for each basis function (i.e., for each value of k2) as the K1 value that minimized the weighted squared difference between model and data. Of the set of 50, the basis function with the smallest residual was selected, and VT was computed as K1/ k2. BP and V3⬙ parametric maps were then created for each subject by applying Equations 4 and 5 to each voxel, where VT(CER) was the mean of all the cerebellar voxelwise estimates for that subject.
(2)
Total distribution volume (VT, mL of plasma gm⫺1 of tissue) was obtained as
V T ⫽ K 1/k 2
(3)
Two measures of SERT availability were calculated. First, the binding potential (BP, mL g⫺1) was derived as the difference between VT in the ROI or voxel of interest (VOI) and VT in the cerebellum, given the low to negligible density of SERT in this region (Ba¨ ckstro¨ m et al 1989; Cortes et al 1988; Laruelle et al 1988):
BP ⫽ V T (ROI or VOI) ⫺ V T(CER)
(4)
Second, the equilibrium specific to nonspecific partition coefficient (V3⬙, unitless) was derived as the ratio of BP to cerebellar VT:
V 3⬙ ⫽ BP/V T(CER)
(5)
The BP and V3⬙ are related to transporter parameters as follows (Laruelle et al 1994):
BP ⫽ f 1B max/K D
(6)
V 3⬙ ⫽ f 2B max/K D ⫽ f 1B max/K DV T(CER)
(7)
where Bmax is the regional concentration of SERT (pmoles g⫺1 of tissue), KD is the affinity of [11C]McN 5652 for SERT (nmol/L), f1is the free fraction in the plasma, and f2 is the free fraction in the nondisplaceable compartment.
Image Analysis PREPROCESSING. Image analysis was performed blind to diagnosis using MEDx (Sensor Systems, Sterling, Virginia). To correct for head movement, PET frames were realigned to a frame of reference using automated image registration (Woods et al 1992) and coregistered to the MRI. ROI ANALYSIS. The ROI boundaries and one region of reference (cerebellum) were drawn on each subject’s MRI; ROIs included dorsal caudate, dorsal putamen, ventral striatum, midbrain, thalamus, hippocampus, amygdala, and anterior cingulate cortex. The midbrain includes various structures with dense SERT concentration (raphe nuclei, substantia nigra, locus ceruleus, ventral tegmental area, and superior and inferior colliculi). Mawlawi et al (2001) describe and illustrate the criteria for delineating the boundaries of striatal substructures (dorsal caudate, dorsal putamen, and ventral striatum, the latter including nucleus accumbens, ventral caudate, and ventral putamen) as well the ability of quantifying radioactive signal from these substructures. Criteria for delineating the other ROIs are available upon request. Right and left ROI volumes were summed.
Statistical Analysis ROI ANALYSIS. Volume, K1, VT, BP, and V3⬙ were each analyzed with univariate repeated-measures analysis of variance (ANOVA); brain region (ROIs and cerebellum) was the withinsubjects factor, and diagnostic group was the between-subjects factor. This model was used because multiple brain regions were examined in the same subject, and it can test for correlations between repeated measures. When the tests of sphericity were significant, the Huynh–Feldt correction for the degrees of freedom was used. Exploratory analyses also compared OCD subjects and control subjects on each parameter within each brain region using two sample unpaired t tests. Within the OCD group, the relationship between SERT availability and OCD severity (measured by the Y-BOCS) was examined with the Pearson’s Product–Moment correlation. All statistical tests were two-tailed with level of significance ␣ ⫽ .05. VOXELWISE ANALYSIS. All voxelwise analyses were performed in the Statistical Parametric Mapping 99 (SPM, Wellcome Department of Cognitive Neurology, University College, London, United Kingdom) environment. The MRIs were normalized to the Montreal Neurologic Institute T1 template. The same transformation was applied to the parametric maps, yielding parametric data normalized into a common template space. These data were smoothed with a Gaussian kernel (FWHM ⫽ 6 mm) to comply with the random fields smoothness assumption employed by the SPM multiple comparisons correction procedure. Two sample unpaired t tests were used to test for group differences in K1, VT, BP, and V3⬙ maps (voxel and cluster significance level p ⬍ .05, corrected for multiple comparisons). Within the OCD group, the linear correlation between OCD severity (measured by the Y-BOCS) and voxelwise BP and V3⬙ was examined.
Serotonin Transporters in OCD
Table 2. Regional Volumes (mm3)
Table 1. Demographic and Clinical Characteristics Parameter Number Mean Age in Years (SD) Number of Men/Women Ethnicity History of SRI Treatment Response to SRI Mean weeks (SD) since last SRI dose Mean Age of OCD Onset in Years (SD) Mean Y-BOCS (SD) Mean HAM-D (SD) DSM-IV Axis I Psychiatric Comorbidity No current or past History of MDD only Subthreshold trichotillomania and history of MDD Both MDD and PTSD Social phobia and history of anorexia History of cannabis abuse History of specific phobia Current/History of Isolated Tics
1417
BIOL PSYCHIATRY 2003;54:1414 –1421
Comparison Subjects
OCD Patients
11 11 31 (10) 31 (12) 6/5 6/5 4AA/5Cau/2H 5AA/5Cau/1H — 6 of 11 — 5 of 6 — 65 (73) — 17 (6) — 20 (4) — 6 (4) 11 — —
3 3 1
— — — — —
1 1 1 1 2/1
AA, African American; Cau, Caucasian; HAM-D, Hamilton Depression Rating Scale; H, Hispanic; SRI, serotonin reuptake inhibitor; MDD, major depressive disorder; OCD, obsessive-compulsive disorder; PTSD, posttraumatic stress disorder; Y-BOCS, Yale–Brown Obsessive–Compulsive Scale.
Results Sample Characteristics More than 120 OCD subjects were screened. Most were ineligible because they were on psychiatric medication or had an excluded current comorbid psychiatric condition (e.g., major depressive disorder (MDD), other excluded anxiety disorder, substance or alcohol abuse). Of the 34 eligible OCD subjects, 23 declined participation, primarily because of concerns about the arterial line and radiation exposure. Demographic and clinical characteristics of the 11 comparison and 11 OCD subjects who participated are shown in Table 1. There were no significant group differences in mean age (t ⫽ .00, df ⫽ 20, p ⫽ 1.00), gender distribution (Fisher’s Exact Test, p ⫽ 1.00), or ethnicity (Fisher’s Exact Test, p ⫽ 1.00). As shown in Table 1, OCD subjects had on average moderate OCD severity: one had mild OCD (Y-BOCS ⫽ 12); most had moderate OCD severity (Y-BOCS ⫽ 16 –23); two had more severe OCD (Y-BOCS ⫽ 24 –29); and none had extreme OCD (Y-BOCS ⱖ32). Although one subject had predominantly symmetry and exactness obsessions with ordering or arranging compulsions, the other 10 endorsed at least two categories of obsessions and two categories of compulsions on the Y-BOCS Symptom Checklist. Within the five-factor model of symptom subtypes (Mataix-Cols et al 1999), these 10 subjects endorsed
Region Cerebellum Dorsal Caudate Dorsal Putamen Ventral Striatum Midbrain Thalamus Hippocampus Amygdala Anterior Cingulate
Comparison Subjects
OCD Subjects
t
p
64,813 (7890) 5519 (595) 8430 (1260) 2042 (654) 6396 (616) 8391 (1698) 7416 (983) 3598 (586) 7101 (1767)
62,750 (13,233) 5300 (1093) 9343 (1865) 2688 (1125) 6782 (844) 8384 (1144) 8019 (1418) 4590 (1188) 7203 (1673)
.44 .58 1.35 1.65 1.23 .01 1.16 2.48 .14
.66 .57 .19 .11 .23 .99 .26 .02 .89
Values are mean (SD), n ⫽ 11 and 11 for comparison and OCD subjects. t and p are value and probability of unpaired two sample t tests (df ⫽ 20), respectively. Right and left volumes were summed. OCD, obsessive– compulsive disorder.
as their current major problem different combinations of the following four factors: contamination obsessions and washing compulsions (n ⫽ 7); aggressive obsessions and checking compulsions (n ⫽ 5); symmetry and exactness obsessions and ordering or arranging compulsions (n ⫽ 7); hoarding obsessions and compulsions (n ⫽ 7). Seven also had prominent mental rituals. Five OCD subjects were SRI naive; the other six had been off SRI medication for an average of 65 weeks (SD ⫽ 73; range: 6 –188 weeks). Three reported infrequent involuntary motor movements that were consistent with isolated motor tics (current, n ⫽ 2; past, n ⫽ 1). None met current or past DSM-IV criteria for a tic disorder.
Injected Dose and Input Function There were no significant group differences in mean injected dose (control subjects, 11.28 ⫾ 4.86 mCi; OCD subjects, 13.51 ⫾ 3.50 mCi; t ⫽ 1.23, df ⫽ 20, p ⫽ .23), mean specific activity at time of injection (control subjects, 880 ⫾ 695 Ci/mmoles; OCD subjects, 834 ⫾ 349 Ci/mmol; t ⫽ .20, df ⫽ 20, p ⫽ .72), injected mass (control subjects, 4.59 ⫾ 1.51 g; OCD subjects, 5.07 ⫾ .91 g; t ⫽ .89, df ⫽ 20, p ⫽ .38) or rate of plasma clearance (control subjects, 147 ⫾ 39 L h⫺1; OCD subjects, 113 ⫾ 52 L h⫺1; t ⫽ 1.72, df ⫽ 20, p ⫽ .10).
Regional Volumes Volumes of ROIs are presented in Table 2. Repeatedmeasures ANOVA showed a significant effect of region (F ⫽ 557, df ⫽ 1.23, 24.55, p ⬍ .001), no effect of group (F ⫽ .083, df ⫽ 1, 20, p ⫽ .78), and no group-by-region interaction (F ⫽ .32, df ⫽ 1.23, 24.55, p ⫽ .62). Regionwise comparisons showed no statistically significant group differences in ROI volumes, with one exception: OCD subjects had significantly larger amygdala than control subjects, a difference that did not survive Bonferroni correction (n ⫽ 9 regions).
1418
H.B. Simpson et al
BIOL PSYCHIATRY 2003;54:1414 –1421
Table 3. Regional Delivery (K1, mL g⫺1 min⫺1) Region Cerebellum Dorsal Caudate Dorsal Putamen Ventral Striatum Midbrain Thalamus Hippocampus Amygdala Anterior Cingulate
Comparison Subjects
OCD Subjects
t
p
.30 (.04) .32 (.04) .36 (.05) .31 (.05) .26 (.03) .33 (.05) .24 (.04) .22 (.03) .32 (.05)
.33 (.08) .34 (.07) .39 (.08) .35 (.08) .27 (.06) .35 (.07) .24 (.04) .26 (.05) .36 (.08)
1.14 .95 1.22 1.34 .51 .69 .70 .98 1.33
.27 .36 .24 .20 .61 .50 .49 .34 .20
Values are mean (SD), n ⫽ 11 and 11 for comparison and OCD subjects. t and p are value and probability of two sample t tests (df ⫽ 20), respectively. OCD, obsessive– compulsive disorder.
ROI Kinetic Analysis Kinetic analysis of time activity curves converged in all cases. K1 values are listed in Table 3. Repeated-measures ANOVA showed a significant effect of region (F ⫽ 140, df ⫽ 5.14, 102.86, p ⬍ .001), no effect of group (F ⫽ 1.11, df ⫽ 1, 20, p ⫽ .31), and no significant group-by-region interaction (F ⫽ 1.79, df ⫽ 5.14, 102.86, p ⫽ .12). Regionwise comparisons revealed no significant group differences in K1 values in any ROI. The VT values are presented in Figure 1. Repeatedmeasures ANOVA showed a significant effect of region (F ⫽ 90, df ⫽ 2.27, 45.35, p ⬍ .001), no effect of group (F ⫽ .04, df ⫽ 1, 20, p ⫽ .85), and no significant group-by-region interaction (F ⫽ .11, df ⫽ 2.27, 45.35, p ⫽ .92). No group differences were observed in cerebellar VT (VT(CER); control subjects: 18.66 ⫾ 3.06 mL g⫺1; OCD subjects: 19.26 ⫾ 3.97 mL g⫺1, t ⫽ .39, df ⫽ 20, p ⫽ .70).
Regional SERT Availability The BP values are presented in Table 4. Repeatedmeasures ANOVA showed a significant effect of region (F ⫽ 73, df ⫽ 2.56, 51.27, p ⬍ .001), no effect of group (F ⫽ .003, df ⫽ 1, 20, p ⫽ .96), and no significant group-by-region interaction (F ⫽ .13, df ⫽ 2.56, 51.27, p ⫽ .92). Regionwise comparisons revealed no significant group differences in BP values in any ROI. The V3⬙ values are presented in Table 5. Repeatedmeasures ANOVA showed a significant effect of region (F ⫽ 108, df ⫽ 3.52, 70.59, p ⬍ .001), no effect of group (F ⫽ .29, df ⫽ 1, 20, p ⫽ .60), and no significant group-by-region interaction (F ⫽ .30, df ⫽ 3.53, 70.59, p ⫽ .86). No significant between-group differences in V3⬙ values were found in the ROIs. There were no significant correlations between OCD severity and [11C]McN 5652 BP or V3⬙ in any ROI. Moreover, post-hoc analyses found no significant mean differences (Mann–Whitney U, all p values ⬎ .25) in ROI
Figure 1. Mean [11C]McN 5652 total distribution volumes (VT, mL g⫺1) and standard deviations in comparison subjects (n ⫽ 11) and subjects with obsessive– compulsive disorder (OCD) (n ⫽ 11) in cerebellum (CER) and regions of interest. Repeatedmeasures analysis of variance showed a significant effect of region (F ⫽ 90, df ⫽ 2.27, 45.35, p ⬍ .001), no effect of group (F ⫽ .04, df ⫽ 1, 20, p ⫽ .85), and no significant group-byregion interaction (F ⫽ .11, df ⫽ 2.27, 45.35, p ⫽ .92). DCA, dorsal caudate; DPU, dorsal putamen; VST, ventral striatum; MID, midbrain; THA, thalamus; AMY, amygdala; HIP, hippocampus; CIN, anterior cingulate.
BP or V3⬙ in OCD patients who were SRI naive (n ⫽ 5) or not (n ⫽ 6) or between OCD patients with a history of tics (n ⫽ 3) and those without (n ⫽ 8).
Voxelwise Analysis Mean K1 maps are shown in Figure 2. In the SPM analysis of the K1 maps, no voxels nor clusters of voxels were Table 4. Regional Binding Potentials (BP, mL g⫺1) Region
Comparison Subjects
OCD Subjects
t
p
Dorsal Caudate Dorsal Putamen Ventral Striatum Midbrain Thalamus Hippocampus Amygdala Anterior Cingulate
18.03 (6.60) 21.78 (6.28) 22.78 (5.96) 29.45 (8.08) 17.37 (4.50) 6.90 (2.23) 20.00 (7.73) 4.23 (1.78)
17.06 (9.16) 23.36 (10.84) 22.64 (10.10) 29.70 (15.19) 17.45 (6.88) 6.97 (3.32) 19.91 (11.00) 4.60 (2.15)
.29 .42 .04 .05 .03 .06 .02 .44
.78 .68 .97 .96 .97 .95 .98 .66
Values are mean (SD), n ⫽ 11 and 11 for comparison and OCD subjects. t and p are value and probability of unpaired two sample t tests (df ⫽ 20), respectively. OCD, obsessive– compulsive disorder.
Serotonin Transporters in OCD
BIOL PSYCHIATRY 2003;54:1414 –1421
1419
Table 5. Equilibrium Specific to Nonspecific Partition Coefficient (V3⬙, unitless) Region Dorsal Caudate Dorsal Putamen Ventral Striatum Midbrain Thalamus Hippocampus Amygdala Anterior Cingulate
Comparison Subjects
OCD Subjects
t
p
.95 (.24) 1.15 (.22) 1.23 (.29) 1.57 (.34) .92 (.17) .37 (.09) 1.04 (.25) .23 (.08)
.84 (.28) 1.17 (.32) 1.14 (.32) 1.49 (.52) .88 (.21) .35 (.13) .99 (.37) .23 (.08)
.90 .12 .70 .44 .54 .27 .43 .23
.38 .90 .49 .67 .60 .79 .67 .82
Values are mean (SD), n ⫽ 11 and 11 for comparison and OCD subjects. t and p are value and probability of unpaired two sample t tests (df ⫽ 20), respectively. OCD, obsessive-compulsive disorder.
significantly different (at p ⬍ .05, corrected for multiple comparisons) between the groups. The SPM analysis of VT maps was negative as well. Mean BP maps are shown in Figure 3. High SERT availability was visualized in midbrain, thalamus, and striatum, in agreement with the ROI analysis and the known distribution of SERT in the human brain (Ba¨ ckstro¨ m et al 1989; Cortes et al 1988; Laruelle et al 1988). In the SPM analysis of the BP maps, no voxels or clusters of voxels were significantly different (at p ⬍ .05, corrected) between the groups. Post hoc analyses of the striatum
Figure 3. Mean parametric maps of binding potential (BP, mL g⫺1) in comparison subjects (n ⫽ 11, top row) and subjects with obsessive– compulsive (OCD) (n ⫽ 11, middle row). Individual BP maps were created as the difference between voxelwise VT and cerebellum VT (i.e., the mean of all the cerebellar voxelwise VT estimates for that subject). Individual BP maps were then spatially normalized to the Montreal Neurologic Institute template (bottom row), and mean BP values for each group were calculated. The BP maps reveal brain areas associated with high serotonin transporter density (i.e., midbrain, thalamus, striatum). No voxels or clusters of voxels were significantly different between the groups in the Statistical Parametric Mapping 99 analysis of these BP maps.
alone (encompassing the three striatal ROIs: dorsal caudate, dorsal putamen, and ventral striatum) demonstrated no voxels nor clusters of voxels significantly different (at p ⬍ .01, uncorrected) between the two groups. Mean V3⬙ maps showed the same pattern of SERT distribution as the BP maps, and the SPM analysis of V3⬙ maps was negative as well. There were no significant correlations between OCD severity and voxelwise BP or V3⬙.
Discussion
Figure 2. Mean parametric maps of tracer delivery to the brain (K1, mL g⫺1 min⫺1) in comparison subjects (n ⫽ 11, top row) and subjects with obsessive-compulsive disorder (OCD) (n ⫽ 11, middle row). Individual K1 maps were derived and spatially normalized to the Montreal Neurologic Institute template (bottom row), and mean K1 values for each group were calculated. No voxels or clusters of voxels were significantly different between the groups in the Statistical Parametric Mapping 99 analysis of these K1 maps. OCD, obsessive-compulsive disorder.
This study compared SERT availability in the brains of subjects with OCD and healthy subjects. Contrary to the hypothesis, there were no significant group differences in SERT availability in the caudate nucleus or any of the ROIs. There also were no significant group differences in SERT binding in the SPM analyses. Within the OCD group, OCD severity and SERT availability were not significantly correlated in either the ROI or the voxelwise analyses. This study had several strengths. First, measurement of the arterial input function enabled the quantitative derivation of distribution volumes, from which both BP and V3⬙
1420
H.B. Simpson et al
BIOL PSYCHIATRY 2003;54:1414 –1421
were calculated. Although methods for deriving [11C]McN 5652 V3⬙ without arterial sampling have been proposed (Ichimiya et al 2002), only with the arterial input function could we demonstrate the absence of group differences in cerebellum distribution volumes and thus validate the use of V3⬙ for between-group comparisons of receptor parameters. Second, ROI and voxel-based analyses were conducted. The ROI method has the advantage of being hypothesis driven, but its ability to detect between-group differences in parameters of interest is limited to differences occurring in anatomically predefined regions. In contrast, the voxelwise approach enables the detection of differences in spatial clusters that include only a fraction of an ROI or that overlap different ROIs. In this study, the absence of SERT alterations in the ROIs of OCD subjects was complemented by an equally negative voxelwise analysis. Third, the OCD sample studied was relatively uncontaminated by SRI treatment: five were SRI naive, and the other six had been free of SRIs for an average of 65 weeks. The study also had several limitations. First, the sample was relatively small (11 OCD and 11 comparisons subjects). The between-subject variability (coefficient of variation) in SERT V3⬙ in the eight ROIs was 24 ⫾ 6% in comparison subjects and 30 ⫾ 7% in OCD subjects, providing adequate power (⬎.80) to detect only relatively large differences in SERT availability. For example, in the dorsal caudate, a decrease of 31% or more in V3⬙ in OCD subjects was required to detect a significant group difference. Nonetheless, there was no trend toward betweengroup differences in the entire sample. The largest difference observed, an 11% decrease in dorsal caudate V3⬙ in OCD subjects, had an effect size of –.38; more than 85 subjects in each group would be needed for this difference to reach significance at the p ⬍ .05 level. On the other hand, we note that one OCD subject, who was not an outlier with respect to the entire sample, was an outlier with respect to the OCD sample for some outcome measures in some ROIs (i.e., BP or V3⬙ ⬎ [upper quartile ⫹ 1.5 ⫻ (interquartile range)]). In post-hoc analyses that excluded this subject, the largest observed difference was then an 18% decrease in dorsal caudate V3⬙ in OCD subjects (t ⫽ 1.72, df ⫽ 19, p ⫽ .10) with an effect size of –.72; the p value still exceeded .25 for all other group comparisons of BP and V3⬙. There was no justification to exclude this patient a priori from the analyses based on experimental data, and there were no obvious clinical differences between this subject and the other OCD subjects. Given the clinical heterogeneity of OCD, these negative results do not exclude the possibility of SERT alterations in subtypes of OCD not included in our sample. For example, because of the need for a drug-free interval before the study (coupled with fears about an arterial line and radiation exposure), no subjects with extreme OCD
severity could be enrolled. Moreover, only one subject reported onset of OCD in childhood, none were known to be SRI unresponsive, none had a chronic tic disorder, and most had a range of OCD symptoms including mental compulsions, as is typical of OCD subjects (Foa and Kozak 1995). Thus, a second limitation of this study is that the findings do not rule out alterations in SERT availability confined to specific clinical profiles not included in this study (e.g., extreme OCD, childhood-onset OCD, SRIunresponsive OCD, comorbid chronic tic disorder, pure hoarders). On the other hand, our findings suggest that OCD in general is not characterized by major decreases in SERT availability in the examined brain regions. Future brain SERT studies might consider studying particular subtypes of OCD. Third, because [11C]McN 5652 is affected by high nonspecific binding, SERT measurement in regions of low SERT density, such as the human neocortex, is unreliable with this tracer. Therefore, the ROI analysis was restricted to regions in which the derivation of SERT BP and V3⬙ has been found reliable (Parsey et al 2000). Thus, this study does not exclude the existence of SERT alterations in OCD in the orbitofrontal (or other parts of the) cortex. The recent development of superior SERT PET radiotracers such as [11C]DASB (Wilson et al 2000) and [11C]AFM (Huang et al 2001, 2002) might permit investigations of the neocortex in the future. An additional benefit of these new tracers is that their free fraction can be measured, allowing one to control for this parameter. In conclusion, this study failed to detect marked differences in SERT availability in subcortical and limbic regions in untreated OCD subjects. Further evaluation of SERT availability in the neocortex and in specific OCD subtypes is desirable to complete this assessment. These results do not exclude the existence of altered 5-HT transmission in OCD that is not associated with SERT density alterations, such as abnormalities of 5-HT release or of 5-HT receptors. Thus, investigations of other aspects of the serotonergic system in OCD subjects are warranted.
This work was supported by the Obsessive-Compulsive Foundation and the National Institute of Mental Health (Grant No. K23 MH01907 to HBS). We thank Gordon Frankle, Wendy Garcia, Ingrid Gelbard-Stokes, Elizabeth Hackett, Leyla Khenissi, Shu-Hsing Lin, Kimchung Ngo, Julie Montoya, Chaka Peters, Norm Simpson, and Kris Wolff for excellent technical assistance.
References Abi-Dargham A, Mawlawi O, Lombardo I, Gil R, Martinez D, Huang Y, et al (2002): Prefrontal dopamine D1 receptors and working memory in schizophrenia. J Neurosci 22:3708 – 3719.
Serotonin Transporters in OCD
Ba¨ ckstro¨ m I, Bergstro¨ m M, Marcusson J (1989): High affinity [3H]paroxetine binding to serotonin uptake sites in human brain tissue. Brain Res 486:261–268. Barr LC, Goodman WK, Price LH, McDougle CJ, Charney DS (1992): The serotonin hypothesis of obsessive compulsive disorder: Implications of pharmacologic challenge studies. J Clin Psychiatry 53(suppl):17–28. Buck A, Gucker PM, Schonbachler RD, Arigoni M, Kneifel S, Vollenweider FX, et al (2000): Evaluation of serotonergic transporters using PET and [11C](⫹)McN-5652: Assessment of methods. J Cereb Blood Flow Metab 20:253–262. Cortes R, Soriano E, Pazos A, Probst A, Palacios JM (1988): Autoradiography of antidepressant binding sites in the human brain: Localization using [3H] [3H]imipramine and [3H] [3H]paroxetine. Neuroscience 27:473–496. Foa EB, Kozak MJ (1995): DSM-IV field trial: Obsessivecompulsive disorder. Am J Psychiatry 152:90 –96. Goodman WK, Price LH, Rasmussen SA, Mazure C, Delgado P, Heninger GR, et al (1989b): The Yale–Brown Obsessive Compulsive Scale. II. Validity. Arch Gen Psychiatry 46:1012–1016. Goodman WK, Price LH, Rasmussen SA, Mazure C, Fleischmann RL, Hill CL, et al (1989a): The Yale–Brown Obsessive Compulsive Scale. I. Development, use, and reliability. Arch Gen Psychiatry 46:1006 –1011. Gunn RN, Lammertsma AA, Hume SP, Cunningham VJ (1997): Parametric imaging of ligand-receptor binding in PET using a simplified reference region model. Neuroimage 6:279 –2787. Hamilton M (1960): A rating scale for depression. J Neurol Neurosurg Psychiatry 23:56 – 62. Huang Y, Bae SA, Zhu Z, Guo N, Hwang DR, Laruelle M (2001): Fluorinated analogues of ADAM as new PET radioligands for the serotonin transporter: Synthesis and pharmacological evaluation. J Labelled Compd Radiopharm 44:S18 –S20. Huang Y, Hwang DR, Narendran R, Sudo Y, Chatterjee R, Bae SA, et al (2002): Comparative evaluation in nonhuman primates of five PET radiotracers for imaging the serotonin transporters: [11C]McN 5652, [11C]ADAM, [11C]DASB, [11C]DAPA, and [11C]AFM. J Cereb Blood Flow Metab 22:1377–1398. Ichimiya T, Suhara T, Sudo Y, Okubo Y, Nakayama K, Nankai M, et al (2002): Serotonin transporter binding in patients with mood disorders: A PET study with [11C](⫹)McN5652. Biol Psychiatry 51:715–722. Laruelle M, van Dyck C, Abi-Dargham A, Zea-Ponce Y, Zoghbi SS, Charney DS, et al (1994): Compartmental modeling of iodine-123-iodobenzofuran binding to dopamine D2 receptors in healthy subjects. J Nucl Med 35:743–754. Laruelle M, Vanisberg M, Maloteaux J (1988): Regional and subcellular localization in human brain of [3H]paroxetine
BIOL PSYCHIATRY 2003;54:1414 –1421
1421
binding, a marker of serotonin uptake sites. Biol Psychiatry 24:299 –309. Mataix-Cols D, Rauch SL, Manzo PA, Jenike MA, Baer L (1999): Use of factor-analyzed symptom dimensions to predict outcome with serotonin reuptake inhibitors and placebo in the treatment of obsessive-compulsive disorder. Am J Psychiatry 156:1409 –1416. Mawlawi O, Martinez D, Slifstein M, Broft A, Chatterjee R, Hwang DR, et al (2001): Imaging human mesolimbic dopamine transmission with positron emission tomography: I. Accuracy and precision of D2 receptor parameter measurements in ventral striatum. J Cereb Blood Flow Metab 21:1034 –1057. McCann UD, Szabo Z, Scheffel U, Dannals RF, Ricaurte GA (1998): Positron emission tomographic evidence of toxic effect of MDMA (“Ecstasy”) on brain serotonin neurons in human beings. Lancet 352:1433–1437. Parsey RV, Kegeles LS, Hwang DR, Simpson N, Abi-Dargham A, Mawlawi O, et al (2000): In vivo quantification of brain serotonin transporters in humans using [11C]McN 5652. J Nucl Med 41:1465–1477. Rauch SL, Cora-Locatelli G, Greenberg BD (2002): Pathogenesis of obsessive-compulsive disorder. In: Stein DJ, Hollander E, editors. Textbook of Anxiety Disorders. Washington, DC: American Psychiatric Association, 191–205. Saxena S, Bota RG, Brody AL (2001): Brain-behavior relationships in obsessive-compulsive disorder. Semin Clin Neuropsychiatry 6:82–101. Shank RP, Vaught JL, Pelley KA, Setler PE, McComsey DF, Maryanoff BE (1988): McN-5652: A highly potent inhibitor of serotonin uptake. J Pharmacol Exp Ther 247:1032–1038. Suehiro M, Scheffel U, Dannals RF, Ravert HT, Ricaurte GA, Wagner H Jr (1993): A PET radiotracer for studying serotonin uptake sites: Carbon-11-McN-5652Z. J Nucl Med 34:120 –127. Szabo Z, Kao PF, Scheffel U, Suehiro M, Mathews WB, Ravert HT, et al (1995): Positron emission tomography imaging of serotonin transporters in the human brain using [11C](⫹)McN5652. Synapse 20:37–43. Szabo Z, Scheffel U, Mathews WB, Ravert HT, Szabo K, Kraut M, et al (1999): Kinetic analysis of [11C]McN5652: A serotonin transporter radioligand. J Cereb Blood Flow Metab 19:967–981. Wilson AA, Ginovart N, Schmidt M, Meyer JH, Threlkeld PG, Houle S (2000): Novel radiotracers for imaging the serotonin transporter by positron emission tomography: Synthesis, radiosynthesis, and in vitro and ex vivo evaluation of (11)Clabeled 2-(phenylthio)araalkylamines. J Med Chem 43:3103– 3110. Woods RP, Cherry SR, Mazziotta JC (1992): Rapid automated algorithm for aligning and reslicing PET images. J Comp Assist Tomogr 16:620 –633.