Inflammation in the Neurocircuitry of Obsessive-Compulsive Disorder

Key Points

Question  Is translocator protein distribution volume, a marker of the microglial component of neuroinflammation, elevated in the cortico-striato-thalamo-cortical (CSTC) circuit of obsessive-compulsive disorder (OCD)?

Findings  In this case-control study, translocator protein distribution volume was significantly elevated in the CSTC of participants with OCD compared with healthy control individuals. The Yale-Brown Obsessive Compulsive Scale measure of distress associated with preventing compulsive behaviors was positively correlated with translocator protein distribution volume in the orbitofrontal cortex.

Meaning  The presence of neuroinflammation in the CSTC of OCD argues for extending investigations of serum autoantibodies to the entire CSTC and that immunomodulatory therapies should be investigated in adult OCD.

Importance  For a small percentage of obsessive-compulsive disorder (OCD) cases exhibiting additional neuropsychiatric symptoms, it was proposed that neuroinflammation occurs in the basal ganglia as an autoimmune response to infections. However, it is possible that elevated neuroinflammation, inducible by a diverse range of mechanisms, is important throughout the cortico-striato-thalamo-cortical circuit of OCD. Identifying brain inflammation is possible with the recent advance in positron emission tomography (PET) radioligands that bind to the translocator protein (TSPO). Translocator protein density increases when microglia are activated during neuroinflammation and the TSPO distribution volume (V_T) is an index of TSPO density.

Objective  To determine whether TSPO V_T is elevated in the dorsal caudate, orbitofrontal cortex, thalamus, ventral striatum, dorsal putamen, and anterior cingulate cortex in OCD.

Design, Setting, and Participants  This case-control study was conducted at a tertiary care psychiatric hospital from May 1, 2010, to November 30, 2016. Participants with OCD (n = 20) and age-matched healthy control individuals (n = 20) underwent a fluorine F 18–labeled N-(2-(2-fluoroethoxy)benzyl)-N-(4-phenoxypyridin-3-yl)acetamide PET scan. It is a high-quality second-generation TSPO-binding PET radiotracer. All participants were drug and medication free, nonsmoking, and otherwise healthy.

Main Outcomes and Measures  The TSPO V_T was measured in the dorsal caudate, orbitofrontal cortex, thalamus, ventral striatum, dorsal putamen, and anterior cingulate cortex. Compulsions were assessed with the Yale-Brown Obsessive Compulsive Scale.

Results  In the OCD and healthy groups, the mean (SD) ages were 27.4 (7.1) years and 27.6 (6.6) years, respectively, and 11 (55%) and 8 (40%) were women, respectively. In OCD, TSPO V_T was significantly elevated in these brain regions (mean, 32%; range, 31%-36% except anterior cingulate cortex, 24%; analysis of variance, effect of diagnosis: P < .001 to P = .004). Slightly lower elevations in TSPO V_T (22%-29%) were present in other gray matter regions. The Yale-Brown Obsessive Compulsive Scale measure of distress associated with preventing compulsive behaviors significantly correlated with TSPO V_T in the orbitofrontal cortex (uncorrected Pearson correlation r = 0.62; P = .005).

Conclusions and Relevance  To our knowledge, this is the first study demonstrating inflammation within the neurocircuitry of OCD. The regional distribution of elevated TSPO V_T argues that the autoimmune/neuroinflammatory theories of OCD should extend beyond the basal ganglia to include the cortico-striato-thalamo-cortical circuit. Immunomodulatory therapies should be investigated in adult OCD, rather than solely childhood OCD, particularly in cases with prominent distress when preventing compulsions.

Obsessive-compulsive disorder (OCD) is a debilitating neuropsychiatric disorder, significantly affecting the function of 1% to 2% of adolescents and adults.¹ A third of OCD cases inadequately respond to pharmacotherapies with good evidence for this condition, such as serotonin reuptake inhibitors and clomipramine.^2,3 Some directions, based on models of striatal dopaminergic hyperactivity with stereotypy and abnormal glutamate uptake/regulation in the cortico-striato-thalamo-cortical (CSTC) circuit, have led to some positive clinical trials of augmentation with antipsychotics and N-methyl-d-aspartate receptor antagonists, respectively.^4,5 Although such findings are being incorporated in treatment algorithms,³ more consistently impactful treatment approaches are needed and the major barrier in therapeutic development is the paucity of novel pathological targets identified in the brain of OCD.

An autoimmune pathophysiology has been proposed for several case series of OCD characterized by prepubertal acute onset, episodic course, and concurrent neurological abnormalities such as choreiform movements occurring or exacerbating after exposure to infection.^6,7 This syndrome has been termed pediatric autoimmune neuropsychiatric disorder associated with group A beta-hemolytic streptococcus (GABHS) (PANDAS) or pediatric acute neuropsychiatric syndrome (PANS). Cross reactivity between gangliosides in basal ganglia neurons with the GABHS cell wall is mechanistically implicated in Sydenham chorea, a series of rapid and uncoordinated jerky movements affecting the face, feet, and hands, which may occur within 6 months of the acute GABHS infection,⁸ and there are some reports of antibasal ganglia antibodies in the serum of PANDAS cases.⁹

Although PANDAS and PANS are relevant for only a minority of OCD cases,^10,11 it is possible that the broader state of neuroinflammation and/or autoimmunity, inducible by a more diverse range of mechanisms, is important in OCD. Prevalence of anxiety disorder is strongly elevated, with rates of 30% to 40% in medical conditions such as systemic lupus erthymatosis and multiple sclerosis, for which autoimmunity is an important component of the medical disease.^12,13 Also, specific prevalence rates of OCD in systemic lupus erythematosus and multiple sclerosis are typically several fold higher.^14,15 In addition, OCD specifically is associated with developmental disorders such as Tourette syndrome and tic disorder for which excessive autoimmunity is also implicated.¹⁶ Moreover, induction of neuroinflammation in rodents with lipopolysaccharide administration is associated with anxiety behaviors, such as reduced exploration in the open field test, even when accounting for changes in motor activity.^17,18

Microglial activation, an important component of neuroinflammation, can be measured with translocator protein (TSPO) positron emission tomography (PET) imaging because microglia increase expression of TSPO when they are activated.¹⁹ In a [1-(2-chlorophenyl)-N-methyl-N-(1-methylpropyl]-3-isoquinoline carboxamide ([¹¹C]PK11195) PET study of 17 children with PANDAS and 15 healthy control adults, Kumar et al,²⁰ using a reference tissue approach, found elevated TSPO binding in the basal ganglia but not thalamic regions. These findings cannot be extrapolated as being applicable to OCD because this study investigated PANDAS rather than OCD. Also, this study restricted its regional analyses to the striatum and thalamus and did not investigate any other region implicated in the neurocircuitry of OCD. Given that with the exception of this neuroimaging study in PANDAS, which represents a small percentage of cases,^10,11 there are no postmortem or neuroimaging studies of neuroinflammation in OCD, hence, it is not known as to whether neuroinflammation occurs in OCD.

While [¹¹C]PK11195 was the first TSPO PET radiotracer, recent advances in PET radioligands for TSPO have resulted in a second generation of radiotracers for which the specific binding signal is at least several-fold higher, such as fluorine F 18–labeled N-(2-(2-fluoroethoxy)benzyl)-N-(4-phenoxypyridin-3-yl)acetamide) ([¹⁸F]FEPPA) PET.^21,22 [¹⁸F]FEPPA is an excellent radioligand for TSPO, having high and selective affinity, considerable brain uptake, and good reversibility.^23,24 Moreover, radioactive metabolites are not detected in brain²⁴ and the binding of [¹⁸F]FEPPA is increased during induced inflammation in animals as well as the positive control condition of Alzheimer disease in humans, which is associated with microglial activation in response to amyloid deposition.^25-27

The primary aim of this study is to compare TSPO distribution volume (V_T) in the CSTC circuit of OCD with healthy control individuals. It is hypothesized that TSPO V_T, being elevated in neuroinflammatory states, is greater in OCD. Because the strongest regional convergence across investigations of neurochemical abnormalities in OCD, such as 5-HT_2A, 5-HTT, 5-HT_1B, and mGluR5 receptor binding^28-33 and fluorodeoxyglucose uptake,³⁴ are in the dorsal caudate and the orbitofrontal cortex, these regions are given the highest priority. However, the neuropathology of OCD also implicates a broader group of structures in the CSTC circuit, such as the thalamus, ventral striatum, dorsal putamen, and anterior cingulate cortex,^4,5,33,34 so these regions were given the next highest priority. Also, given that neuroinflammatory pathologies of the CSTC circuit, including vascular disease, tumors, Huntington disease, Tourette disorder, and Sydenham chorea, are associated with disturbances of complex motor behavior, the third aim, which was exploratory, is to assess the association between TSPO V_T in the dorsal caudate and orbitofrontal cortex with severity of compulsions in OCD.

Twenty participants with OCD and 20 age-matched healthy control individuals (within 4 years) completed the study. Participants were recruited from the Toronto, Ontario, Canada, area at a tertiary care psychiatric hospital (Centre for Addiction and Mental Health) from May 1, 2010, to November 30, 2016. All were aged 19 to 48 years, nonsmoking, and in good physical health (see demographic characteristics in Table 1). None of the participants had a history of autoimmune disease, and all were free of medical illnesses for at least 4 weeks. Fourteen healthy control individuals (70%), reported in a previous study,³⁵ were also included in the present study. No other healthy or OCD cases in this study were included in any other study. The remaining control individuals have not been included in other studies. Eleven participants (55%) had OCD prior to age 12 years and 18 participants (90%) had OCD onset prior to age 20 years. The presence or absence of psychiatric disorders and OCD specifically were confirmed using the Structured Clinical Interview of DSM-IV.³⁶ Two OCD cases had a history of a major depressive episode, a single episode occurring 4 and 5 years prior to their scan date. The severity of OCD was measured with the Yale-Brown Obsessive Compulsive Scale (Y-BOCS).^37,38 Clinical symptoms of OCD varied and severity of the obsessive-compulsive symptoms was moderate to severe on average (eTable 1 in the Supplement). None of the OCD cases met criteria for PANDAS or PANS.^6,7

Participant diagnosis was verified by a psychiatrist (J.H.M. or L.R.). Exclusion criteria for all participants were pregnancy; herbal, drug, or medication (including psychotropic) use within the past 6 weeks, except for oral contraceptives; substance abuse (including cigarette smoking); and any history of neurologic illness or injury. Other exclusion criteria consisted of concurrent active Axis I disorders, current or past alcohol or substance dependence, bipolar I or II disorder, and borderline or antisocial personality disorder, the latter ruled out with the Structured Clinical Interview for DSM-IV Axis II disorders.³⁹ All participants underwent urine drug screening, and women received a urine pregnancy test on the screening and PET scanning days. Participants provided written informed consent after all procedures were fully explained. The protocol and informed consent forms were approved by the research ethics board of the Centre for Addiction and Mental Health, Toronto, Ontario, Canada.

Each participant underwent a single [¹⁸F]FEPPA PET scan conducted at the Research Imaging Centre at the Centre for Addiction and Mental Health. Intravenous [¹⁸F]FEPPA²⁴ was administered as a bolus (mean [SD], 183.5 [10.5] MBq or 4.9 [0.3] mCi). The [¹⁸F]FEPPA was of high radiochemical purity (>98.0%) and high specific activity (mean [SD], 133.6 [118.7] TBq/mmol). The PET scan duration was 125 minutes after the injection of [¹⁸F]FEPPA. Positron emission tomographic scans were acquired using a 3-dimensional brain scanner (HRRT; CPS/Siemens). Positron emission tomographic scans were acquired and reconstructed as described previously.^{21,24,35,40,41} All PET images were corrected for attenuation using a single photon point source, cesium 137 (half-life, 30.2 years; energy, 662 keV), and were reconstructed using a filtered back-projection algorithm, with a Hann filter at Nyquist cut-off frequency.⁴¹ Manual and automatic blood sampling (ABSS, Model PBS-101; Veenstra Instruments) was conducted to determine the input function of parent compound in plasma for the kinetic analysis, based on the ratio of radioactivity in whole blood to radioactivity in plasma, and the percentage of parent compound in plasma.²¹ A biexponential function was used to fit the blood-to-plasma ratios and a Hill function was used to fit the percentage of unmetabolized tracer as previously described.^21,35 A 2-tissue compartment model was applied to generate time activity curves from each region of interest (ROI) to quantitate TSPO V_T with [¹⁸F]FEPPA PET, as this was previously demonstrated to be the optimal model for [¹⁸F]FEPPA PET.²¹

Regions of interest were automatically generated using the semiautomated software (ROMI; Research Imaging Centre, Centre for Addiction and Mental Health). ROMI is based on the individualization of a set of standard ROIs using magnetic resonance imaging (MRI) coregistered with the PET image, followed by a step of gray matter voxel selection, which incorporates the probability of gray matter based on the segmentation of the individual MRI as previously described.^42,43 For the anatomical delineation of ROIs, the first 13 participants (OCD: n = 8; healthy: n = 5) underwent 2-dimensional axial proton density MRI acquired with a General Electric Signa 1.5-T MRI scanner (section thickness, 2 mm; repetition time, >5300 milliseconds; echo time, 13 milliseconds; flip angle, 90°; number of excitations, 2; acquisition matrix, 256 × 256; and field of view, 22 mm). The remaining 27 participants (OCD: n = 12; healthy: n = 15) underwent 2-dimensional axial proton density MRI acquired with a General Electric Signa 3-T MRI scanner (section thickness, 2 mm; repetition time, 6000 milliseconds; echo time, 8 milliseconds; flip angle, 90°; number of excitations, 1; acquisition matrix, 256 × 192; and field of view, 16.5 mm). A subset of participants were scanned with both MRI machines to rule out bias due to delineation with ROMI. The differences in MRI acquisitions resulted in near identical TSPO V_T values and regional volumes in a sample undergoing both MRI (n = 12; r > 0.97; P < .001; J.H.M., unpublished data, May 2010-February 2014).

The binding affinity of [¹⁸F]FEPPA for TSPO is affected by a single-nucleotide polymorphism (rs6971; C→T) in exon 4 of the TSPO gene (NCBI Entrez Gene 706).^41,44 Individuals with high-affinity binding (Ala147/Ala147) and mixed-affinity binding (Ala147/Thr147) account for more than 95% of the population in the greater Toronto area.^35,45 The polymorphism rs6971was genotyped as described previously.⁴¹ Homozygotes for the low binding gene (Thr147/Thr147) were excluded from data analysis.

To test the primary hypothesis, a multivariate analysis of variance (MANOVA) was applied with TSPO V_T in the dorsal caudate, orbitofrontal cortex, thalamus, ventral striatum, dorsal putamen, and anterior cingulate cortex as the dependent variables, with diagnosis as the key independent predictor variable and genotype (rs6971 polymorphism) as an additional independent variable. Main effects were considered significant at the conventional P ≤ .05. Effects in each region, analyzed by univariate analysis of variance (ANOVA), were considered significant after Bonferroni correction at P ≤ .009.

To assess whether there was a global difference in TSPO V_T between participants with OCD and healthy control individuals, a MANOVA was also applied with TSPO V_T from all regions sampled included as the dependent variables, with diagnosis and genotype as independent predictor factors. The association of TSPO V_T in the dorsal caudate and orbitofrontal cortex with time occupied, interference, distress, resistance, and control over obsessions and compulsions was assessed applying a partial correlation coefficient controlling for genotype. This was treated as a posthoc analysis. Statistical analyses were performed using IBM SPSS version 22.0.⁴⁶

To investigate whether there was evidence for a pattern of TSPO V_T elevation in OCD that differs as compared with the elevations found in major depressive disorder,³⁵ a post hoc analysis was done comparing the relative TSPO V_T values in the dorsal caudate and orbitofrontal cortex, regions most implicated in OCD, in relation to TSPO V_T in the anterior cingulate cortex, a region implicated in major depressive disorder (eAppendix in the Supplement).

Translocator protein V_T was significantly greater in the OCD cases within the dorsal caudate, orbitofrontal cortex, thalamus, ventral striatum, dorsal putamen, and anterior cingulate cortex (MANOVA; main effect of diagnosis: F_6,32 = 2.8, P = .03; main effect of genotype: F_6,32 = 6.2, P < .001) (Figure 1). Differences in individual regions were robust (ANOVAs: dorsal caudate: F_1,37 = 15.5, P < .001; orbitofrontal cortex: F_1,37 = 13.5, P = .001; thalamus: F_1,37 = 13.2, P = .001; ventral striatum: F_1,37 = 14.4, P = .001; dorsal putamen: F_1,37 = 12.8, P = .001; and anterior cingulate cortex: F_1,37 = 9.5, P = .004, with elevations of 35.6%, 30.9%, 33.5%, 33.8%, 32.6%, and 23.5% magnitudes, respectively) (Figure 1 and Table 2). In addition, generally lesser but significant elevations of TSPO V_T were observed in all gray matter regions sampled (Table 2). There was no effect of age or sex on TSPO V_T in any region (MANOVA; main effect of age: F_6,31 = 0.69, P = .66; main effect of sex: F_6,31 = 0.85, P = .54).

Greater TSPO V_T in the orbitofrontal cortex was significantly correlated with greater distress associated with preventing compulsive behaviors as reported on the Y-BOCS (uncorrected Pearson partial correlation coefficient controlling for genotype; r = 0.62; P = .005) (Figure 2 and Table 3).

Reanalysis of the data set omitting the 2 patients who had a history of a major depressive episode demonstrated the results were similarly significant (MANOVA; main effect of diagnosis: F_6,30 = 2.5, P = .04; main effect of genotype: F_6,30 = 5.5, P = .001). Differences in individual regions were robust (ANOVAs; dorsal caudate: F_1,35 = 13.3, P = .001; orbitofrontal cortex: F_1,35 = 13.8, P = .001; thalamus: F_1,35 = 13.8, P = .001; ventral striatum: F_1,35 = 14.0, P = .001; dorsal putamen: F_1,35 = 13.8, P = .001; and anterior cingulate cortex: F_1,35 = 9.3, P = .004). The eFigure and eTable 2 in the Supplement show further results comparing the relative TSPO V_T values across the dorsal caudate and anterior cingulate cortex between those with OCD and those with major depressive episode, as well as an itemized list of Hamilton Depression Rating Scale score results in these samples.

To our knowledge, this is the first study to investigate inflammation in the brain of OCD. The most prominent finding is greater TSPO V_T throughout the CSTC circuit implicated in OCD including the orbitofrontal cortex, dorsal caudate, thalamus, dorsal putamen, ventral striatum, and, albeit to a slightly lesser extent, the anterior cingulate cortex. This finding was also present in the other gray matter regions of brain at a slightly lower level of magnitude and this regional distribution is consistent with a neuroinflammatory phenotype. These results have important implications for the pathophysiology of OCD and novel therapeutic treatment directions.

The best explanation for elevated TSPO V_T is that it reflects neuroinflammation consequent to microglial activation. Across diverse models of cerebral artery occlusion, toxin exposure, and lipopolysaccharide administration, it is consistently demonstrated that the temporal course of elevations in TSPO levels are highly correlated with markers of microglial activation.^47-49 The widespread distribution of elevated TSPO V_T throughout gray matter is consistent with investigations of TSPO V_T with more sensitive radioligands in diseases with microglial activation and more focal pathologies such as Alzheimer disease and stroke, which tend to find that no gray matter region is fully spared from an elevation of TSPO V_T.^26,27,50 This could be accounted for by diffuseness of initiating disease pathology; spread of microglial activating disease–associated molecular patterns and/or pathogen-associated molecular patterns; and most plausibly, paracrine effects of activated microglia.⁵¹ Even so, it is interesting that TSPO V_T was elevated by more than 30% within the orbitofrontal cortex, dorsal caudate, thalamus, dorsal putamen, and ventral striatum. The involvement of the CSTC has important implications for the autoimmune etiology of OCD. Traditionally, investigations of autoantibodies in serum of OCD cases mostly focus on striatum, presumably because GABHS is associated with Sydenham chorea and the PANDAS mechanism is based on the epitope cross reactivity of the striatum with GABHS.^6,7 Some investigators are beginning to evaluate serum autoantibodies to other parts of the CSTC circuit such as the thalamus,⁵² but our study argues that it is critical that such studies consistently pursue autoantibodies for the dorsal caudate and orbitofrontal cortex as well as the remainder of the CSTC circuit.

This study demonstrates that microglial activation occurs well beyond the initial, frequently childhood, onset of OCD, and the presence of activated microglia provides a useful opportunity to modulate their function as a therapeutic strategy. Although pharmaceutical development does not traditionally prioritize OCD, neuromodulatory treatments under development for other diseases associated with microglial activation, such as Alzheimer disease,^53,54 might be repurposed toward OCD. Microglial activation may include components harmful to neurons and glia (termed M1 responses), such as creating reactive oxygen and nitrogen species and producing proinflammatory cytokines, but also include helpful components (termed M2 responses) such as clearing cellular debris, inducing angiogenesis, and promoting tissue repair.⁵⁵ More specifically, the current study would suggest that medications under investigation for shifting activated microglia from an M1 to M2 state, such as azithromycin, minocycline, and bexarotene,⁵⁶ should be investigated in OCD, particularly in samples in which the severity of distress associated with preventing compulsive behaviors is highly prevalent. Consistent with consideration of an immunomodulatory approach, minocycline, an antibiotic that, among its effects, inhibits major histocompatibility complex II expression and reduces M1 type markers,⁵⁷ significantly reduced Y-BOCS scores in a recent randomized, double-blind, placebo-controlled study of add-on to fluvoxamine in 102 patients with OCD.⁵⁸

This study has several limitations, most of which are attributable to the interpretation of TSPO V_T and the application of PET imaging. Elevated TSPO V_T is well-established as a marker of microglial activation; however, given that TSPO has roles in translocating cholesterol from outer to inner cell membranes and may form an oligomer with the mitochondrial permeability transition pore,⁵⁹ it is theoretically possible that as knowledge regarding TSPO increases, other factors will be identified that may influence TSPO binding in the brain. Also, binding of the PET radiotracer can be affected by changes in density and affinity of the target. In addition, the association between elevated TSPO and OCD, as well as greater distress associated with preventing compulsive behaviors, indicates significant associations among these phenomena but does not establish a causal one as this issue will require future study through a combination of longitudinal studies and assessment of immunomodulatory interventions.

In summary, to our knowledge, this study is the strongest evidence to date for inflammation in the brain in OCD. The demonstration of elevated TSPO V_T in the CSTC circuit addresses a critical gap in the autoimmune/neuroinflammatory theory of OCD. The regional distribution of greater TSPO V_T throughout the CSTC circuit argues for consideration of autoimmune mechanisms beyond the basal ganglia and suggests a new opportunity for repurposing immunomodulatory therapies to treat OCD.

Accepted for Publication: May 3, 2017.

Corresponding Author: Jeffrey H. Meyer, MD, PhD, FRCP(C), Research Imaging Centre, Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health, 250 College St, Ste B26, Toronto, ON M5T 1R8, Canada (jeff.meyer@camhpet.ca).

Published Online: June 21, 2017. doi:10.1001/jamapsychiatry.2017.1567

Author Contributions: Dr Meyer had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Meyer.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Attwells, Setiawan, Rusjan, Mizrahi, Meyer.

Critical revision of the manuscript for important intellectual content: Attwells, Setiawan, Wilson, Miler, Xu, Richter, Kahn, Kish, Houle, Ravindran, Meyer.

Statistical analysis: Attwells, Setiawan, Miler, Meyer.

Obtained funding: Houle, Meyer.

Administrative, technical, or material support: Wilson, Xu, Kahn, Houle, Meyer.

Study supervision: Wilson, Mizrahi, Houle, Meyer.

Conflict of Interest Disclosures: Drs Wilson, Houle, and Meyer have received operating grant funds for other studies from Janssen, Bristol-Myers Squibb, Eli Lilly and Company, GlaxoSmithKline, Lundbeck, and SK Life Sciences in the past 5 years. Dr Meyer has been a consultant to Mylan, Lundbeck, Takeda, Teva, and Trius in the past 5 years. None of these companies participated in the design or execution of this study or in the writing of the manuscript. Dr Meyer is an inventor on 4 patents of blood markers to predict brain inflammation and/or to diagnose affective disorders and a dietary supplement to reduce depressed mood in the postpartum period. Dr Meyer is arranging collaborations with nutraceutical companies for the dietary supplement. No other disclosures were reported.

Funding/Support: This study was supported by a Canada Research Chair from the Canadian Institutes of Health Research (Dr Meyer), an operating grant from the Grant Family, and funding from the Canadian Foundation for Innovation and the Ontario Ministry for Research and Innovation (infrastructure).

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

Meeting Presentations: This study was presented in part at the Society of Biological Psychiatry Annual Meeting; May 18, 2016; Atlanta, Georgia; and the American College of Neuropsychopharmacology Annual Meeting; December 6, 2016; Hollywood, Florida.

Additional Contributions: James L. Kennedy, MD (Tanenbaum Centre for Pharmacogenetics), assisted with genotyping; Nathan Kolla, MD, PhD (Research Imaging Centre, Institute of Medical Science, University of Toronto), assisted with medical coverage; Alvina Ng, BSc, and Laura Nguyen, BSc, worked as study technicians; and Jun Parkes, MSc, Armando Garcia, BSc, Winston Stableford, BSc, and Min Wong, BSc (Research Imaging Centre), served as chemistry staff. Also, Terry Bell, BSc, and Ted Harris-Brandts, BSc (Research Imaging Centre), provided engineering support. All contributors are paid employees of the Centre for Addiction and Mental Health.