DRD1 and DRD2 Splice Variant Expression in Schizophrenia and Affective Disorders, and Associations with Snps in Postmortem Brain

ORIGINAL ARTICLE Contrasting changes in DRD1 and DRD2 splice variant expression in schizophrenia and affective disorders, and associations with SNPs in postmortem brain

SS Kaalund1,2,3, EN Newburn1,TYe4,RTao4,CLi4, A Deep-Soboslay4, MM Herman1, TM Hyde4, DR Weinberger4, BK Lipska1,5 and JE Kleinman4,5

Dopamine 2 receptor (DRD2) is of major interest to the pathophysiology of schizophrenia (SCZ) both as a target for antipsychotic drug action as well as a SCZ-associated risk gene. The dopamine 1 receptor (DRD1) is thought to mediate some of the cognitive deﬁcits in SCZ, including impairment of working memory that relies on normal dorsolateral prefrontal cortex (DLPFC) function. To better understand the association of dopamine receptors with SCZ, we studied the expression of three DRD2 splice variants and the DRD1 transcript in DLPFC, hippocampus and caudate nucleus in a large cohort of subjects (~700), including patients with SCZ, affective disorders and nonpsychiatric controls (from 14th gestational week to 85 years of age), and examined genotype-expression associations of 278 single-nucleotide polymorphisms (SNPs) located in or near DRD2 and DRD1 genes. Expression of D2S mRNA and D2S/D2-long (D2L) ratio were signiﬁcantly increased in DLPFC of patients with SCZ relative to controls (Po0.0001 and Po0.0001, respectively), whereas D2L, D2Longer and DRD1 were decreased (Po0.0001). Patients with affective disorders showed an opposite pattern: reduced expression of D2S (major depressive disorder, Po0.0001) and increased expression of D2L and DRD1 (bipolar disorder, Po0.0001). Moreover, SCZ-associated risk alleles at rs1079727, rs1076560 and rs2283265 predicted increased D2S/D2L expression ratio (Po0.05) in control individuals. Our data suggest that altered splicing of DRD2 and expression of DRD1 may constitute a pathophysiological mechanism in risk for SCZ and affective disorders. The association between SCZ risk-associated polymorphism and the ratio of D2S/D2L is consistent with this possibility.

Molecular Psychiatry (2014) 19, 1258–1266; doi:10.1038/mp.2013.165; published online 10 December 2013 Keywords: bipolar disorder; depression; development; dopamine receptors; genotype; schizophrenia

INTRODUCTION translocation of the splice regulator PTBP1.14,15 D2S acts mainly as Schizophrenia (SCZ) is a neuropsychiatric disorder characterized a presynaptic receptor harboring the autoreceptor function of D2 by hallucinations, delusions, apathy, social withdrawal and cogni- receptors, whereas the postsynaptic actions are mediated by tive deficits. Dopamine is one of the neurotransmitters that has the D2L isoform.12,16,17 been implicated in SCZ. The psychotic symptoms and cognitive The interest in the DRD1 gene in SCZ is primarily related to symptoms are thought to be related to dopamine 2 receptor cognition.18 There is, however, some evidence for the changes in (DRD2) and dopamine 1 receptor (DRD1), respectively. The fact the DRD1 expression in SCZ and genetic variation in DRD1 that that neuroleptics are effective in treating psychotic symptoms in increases risk for SCZ.19 The DRD1 gene has no introns and there proportion to their DRD2 blocking ability in striatum1,2 led to the are no known alternative transcripts. hypothesis of increased dopamine signaling in the striatum, which In this study, we sought to confirm and extend the findings has been further supported by a number of in vivo imaging of the associations between genetic variation that increases risk studies.3–7 Recently, it has also been shown that genetic variation for SCZ and expression of specific alternative transcripts, and to in the DRD2 gene is associated with SCZ8,9 but the mechanism of examine diagnostic and neuroanatomical specificity of these this association remains unclear. It may involve alternative associations and neurodevelopmental profiles of expression of transcripts, as the rs2283265 risk allele (G)8 has been associated DRD2 and DRD1 transcripts. For this, we measured the expression with increased mRNA expression of the DRD2 short variant (D2S)10 of DRD2 and DRD1 transcripts in the dorsolateral prefrontal cortex and increased receptor availability in the striatum.11 The three (DLPFC), hippocampus (HPC) and caudate nucleus of the striatum DRD2 variants, D2S, D2-long (D2L) and D2Longer12,13 are in a large cohort of patients with SCZ, bipolar disorder (BPD), generated from alternative splicing of the DRD2 gene, a process major depression disorder (MDD) and nonpsychiatric controls. We that may be dependent on dopamine signaling, where activation also examined the associations between genetic variation and of the receptor increases the expression of D2L through nuclear expression in nonpsychiatric controls. In addition, as SCZ is a

1Human Brain Collection Core, IRP, National Institute of Mental Health, Bethesda, MD, USA; 2Research Laboratory for Stereology and Neuroscience, Bispebjerg University Hospital, Copenhagen NV, Denmark; 3 Faculty of Health Sciences, Protein Laboratory, Institute of Neuroscience and Pharmacology, University of Copenhagen, Copenhagen, Denmark and 4Lieber Institute for Brain Development, Baltimore, MD, USA. Correspondence: Dr BK Lipska, Human Brain Collection Core, IRP, National Institute of Mental Health, Bethesda, MD 20892, USA. E-mail: [email protected] 5These authors contributed equally to this work. Received 6 June 2013; revised 4 October 2013; accepted 17 October 2013; published online 10 December 2013 DRD1 and DRD2 expression in psychiatric disorders SS Kaalund et al 1259 neurodevelopmental disorder in which abnormal dopamine for APDs was seen in 62.5% (n = 110) of the SCZ patients and signaling and aberrant alternative splicing during brain develop- 29.8% (n = 17) of the BPD. ment have been suggested,20–22 we examined temporal expression trajectories of DRD1 and DRD2 transcripts in DLPFC and HPC DNA collection and genotyping across the human life span. DNA was extracted from cerebellar tissues (Qiagen, Valencia, CA, USA). All brain samples were genotyped using Illumina Human MATERIALS AND METHODS 650K or 1M-Duo BeadChips (Illumina, San Diego, CA, USA) according to the manufacturer’s instructions. Genotype accuracy Human postmortem tissue was assessed by regenotyping within a subsample, and reprodu- Postmortem brains were collected in the Section on Neuropathol- cibility was routinely >0.99. In this study, we examined 278 single- ogy of the Clinical Brain Disorder Branch at the National Institute of nucleotide polymorphisms (SNPs) located in or near DRD1 and Health with informed consent of the next kin under National DRD2 genes (Supplementary Tables S1 and S2). Institute of Mental Health protocol 90-M-0142 and from the National Institute of Child Health and Human Development Brain and Tissue RNA extraction and quantitative real-time PCR Bank for Developmental Disorders under contracts NO1-HD-4-3368 Total RNA was extracted from ~100 mg of tissue using the RNeasy and NO1-HD-4-3383. Clinical characterization, neuropathological ’ screening, toxicological analyses and dissections of the DLPFC, kit (Qiagen) according to the manufacturer s protocol. The yield of HPC and caudate were performed as previously described.23,24 total RNA was determined by spectophotometry by measuring the A total of 701 DLPFC samples from postmortem brains were used absorbance at 260 nm. The RNA quality was assessed by high- for this study including 148 specimens collected through the resolution capillary electrophoresis on an Agilent Bioanalyzer 2100 Stanley Medical Research Institute. The set of the samples from (Agilent Technologies, Palo Alto, CA, USA), samples with RNA integrity number (RIN)o5 were excluded. cDNA was created from nonpsychiatric individuals consists of 326 subjects, ranging from μ fetal weeks 14–20 (N = 43) to 85 years of age. Diagnostic studies 4 g total RNA using SuperScript First-Strand Synthesis System for real-time PCR (Invitrogen, Carlsbad, CA, USA), according to the were carried out using 176 subjects with SCZ, 61 subjects with BPD, ’ 24,25 138 subjects with MDD and 244 controls (the youngest patient manufacturer s protocol. We and others recognize the general being 13 years of age). Demographic data for these samples are importance of RNA quality for this type of studies, and have extensively researched and published on the relationships summarized in Table 1. To address the question of regional 23 speciﬁcity, 451 HPC samples and 177 caudate samples, which between RIN and pH and gene expression. Based on these data, we chose to include only samples with RIN ⩾ 5 in all our overlapped with the DLPFC samples (96% and 94%, respectively), o were also used in this study. Caudate samples did not include recent studies. Samples with lower quality, which constitute 5% subjects with MDD and were all from subjects >13 years of age. of all RNA samples extracted in our laboratory, are not included in any quantitative PCR or microarray assays. Regardless, we used RIN as covariates in all our analyses. Antemortem medication Expression levels of mRNA were measured using an ABI Postmortem toxicology screening was performed by the medical Prizm 7900 sequence detection system with 384-well format examiner on every sample to test for illicit drug use. Prescription (Applied Biosystems, Foster City, CA, USA) using Taqman assays- drug use at time of death, including antipsychotic (APD) and by-design and custom-made assays designed using PRIMER antidepressant medications, was also measured in postmortem EXPRESS software (Applied Biosystems; D2S Hs_01014210_m1, blood and/or cerebellar tissue by the medical examiner and/or by DRD1 Hs_00265245_s1, D2L Forward primer: 3′-TGCACCGTTATCA National Medical Services in order to determine which prescribed TGAAGTCTAATG-5′, Reverse primer: 3′-CGGGCAGCCTCCACTCT-5′, medications were being used at the time of death. Positive results Probe: 3′-AGTTTCCCAGTGAACAGG-5′, D2Longer Forward primer:

Table 1. Demographic summary of postmortem brain samples

Diagnosis n Sex F/M Age mean in years Race (Cauc/AA/Other) pH RIN

DLPFC Control (>13 years) 244 73/171 40 111/121/12 6.5 8.3 Bipolar disorder 61 25/36 45 51/6/4 6.4 8 Major depression 138 59/36 45 119/14/2 6.4 8 Schizophrenia 176 65/111 50 96/73/7 6.4 7.8 Control (fetal) 43 21/22 14–20 week 5/37/1 NA 8.8 Control (0–13 years) 39 12/27 3 21/18/1 6.4 7.9

Hippocampus Control (>13 years) 192 69/128 40 77/107/8 6.5 8 Bipolar 31 10/21 46 27/2/2 6.2 7.8 MDD 72 36/36 46 56/13/3 6.2 7.8 Schizophrenia 101 36/65 52 40/56/5 6.3 7.4 Control (fetal) 29 16/13 14–20 week 3/25/1 NA 9.5 Control (0–13 years) 26 10/16 3 14/12/0 6.4 7.5

Caudate nucleus Control (>13 years) 78 26/52 41 32/43/2/1 6.6 8.1 Bipolar 44 14/30 43 38/3/3/1 6.3 8 Schizophrenia 55 17/38 53 21/32/0/2 6.4 7.8 Abbreviations: AA, African–American; Cauc, Caucasian; DLPFC, dorsolateral prefrontal cortex; F, female; M, male, MDD, major depression disorder; n, number of samples; NA, not applicable; Other: Hispanic and Asian American; pH, cerebellar pH; RIN, RNA integrity number.

© 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 1258 – 1266 DRD1 and DRD2 expression in psychiatric disorders SS Kaalund et al 1260 3′-GGACATGAAACTCTGCACCGTTA-5′, Reverse primer: 3′-GCAC Table 2. Main effect of diagnosis on expression by analysis of CACTCTCCGCCTGTT-5′, Probe: 3′-CATGAAGTCTAATGGGAGTTT covariance -5′). The PCR data were acquired from the Sequence Detector fi Software (SDS version 2.0, Applied Biosystems) and quanti ed by a Brain region Transcript df F P standard curve method as described previously.23 Expression data were normalized to a geometric mean of three housekeeping DLPFC D2S (3, 610) 30.59 o0.0001 genes such as ACTB, GUSB and B2M. D2L (3, 610) 80.43 o0.0001 D2S/D2L (3, 610) 149.31 o0.0001 D2Longer (3, 610) 15.37 o0.0001 Statistical analysis D1 (3, 613) 75.17 o0.0001 The data were analyzed using the statistical software program R Hippocampus D2S (3, 388) 4.36 o0.01 (version 2.12.0). The distribution of expression values were plotted D2L (2, 387) 7.56 o0.0001 for each transcript individually, and based on these distribution D2S/D2L (3, 386) 5.63 o0.0001 o plots the data were log2 transformed to obtain a normal D2Longer (3, 393) 3.30 0.05 D1 (3, 396) 1.46 NS (Gaussian) distribution. The effects of diagnosis on expression Caudate D2S (2, 171) 1.01 NS were analyzed individually for each transcript using General Linear D2L (2, 179) 0.33 NS Model univariate analysis, that is, analysis of covariance. This D2S/D2L (2, 172) 0.48 NS approach was chosen because some predictive variables were D1 (2, 170) 2.33 NS categorical (for example, sex) and some were continuous (for example, age). We chose to analyze the data as univariate instead Abbreviations: DLPFC, dorsolateral prefrontal cortex ; D1, dopamine 1; D2, dopamine 2; D2L,D2-long ; NS, not significant. Covariates included age, of multivariate (that is, all transcripts at once) because the race and RNA integrity number; P>0.05. expression of each transcript was measured individually and not in a competitive real-time PCR assay. Because the dependent variable (mRNA expression) is continuous, normally distributed and we assumed equal variance, we used analysis of covar- o iance with age, sex (female and male), race (Caucasian, African- greater in patients with SCZ (P 0.0001) compared with patients with BPD and MDD. Expression of D2L (Po0.0001) and D2Longer American, Asian, Hispanic), postmortem interval, pH and RIN as o covariates. Outliers were identified by the Bonferoni outlier test (P 0.05) was lower in patients with MDD relative to BPD patients. The ratio of D2S to D2L was greater in patients with MDD and removed. The number of outliers ranged from 1 to 6. Tukey o honest significant difference post hoc comparisons were used to (P 0.01) compared with BPD patients. fl In the HPC, the main effects of diagnosis on expression reached evaluate diagnostic group differences. The in uence of age on fi gene expression in normal controls across the life span was signi cance for D2S, D2L and D2Longer, but not for DRD1 (Table 2). The results of post hoc tests demonstrated that D2S expression examined by multiple regression with age as a continuous fi o predictor, and sex (male or female), race (individuals of European was signi cantly reduced in SCZ (P 0.01) as compared with descent or African–Americans) and developmental stage (fetal controls, that is, the direction of change was opposite to that – observed in the DLPFC. BPD patients showed significantly period and postnatal periods (0 20 years and ages older than o o 20 years)) were categorical predictors. The use of age, stage and increased levels of D2L (P 0.01) and D2Longer (P 0.05) in the age-by-stage interaction allowed the effects of age on gene HPC, similarly to the changes observed in the DLPFC. There were expression to be independent in each age stage (that is, to have no differences between MDD patients and controls for any of the fi transcripts examined in the HPC. Patients with SCZ had different slopes between the stages), and thus enabled tting fi o o nonlinear trajectories over the entire life span into linear models signi cantly less D2S (P 0.05), D2L (P 0.0001) and D2Longer fi (Po0.05) compared with patients with BPD. Patients with BPD within life stages. The data are presented in gures using LOESS fi o fits. For examining correlations between expression of transcripts, had signi cantly less D2L (P 0.0001) than patients with MDD. we used Pearson’s correlation coefficient. There were no differences between groups of D2S to D2L ratio. In the caudate, there were no significant effects of diagnosis on any of the DRD2 or DRD1 transcripts (Table 2). RESULTS Even though the differential expression changes observed for patients with SCZ and BPD (some of whom have also been treated Changes in DRD1 and DRD2 gene expression in SCZ and affective with APDs) displayed regional and diagnostic specificity, APD med- disorders ication may have contributed to the observed effects. To address In the DLPFC, there were significant main effects of diagnosis on this issue, we inquired whether expression differed between expression of all transcripts examined, D2S, D2L, D2S/D2L, patients tested positive and negative for APDs at the time of D2Longer and DRD1 (Table 2). Post hoc analyses indicated that death. We did not observe any significant differences in in patients with SCZ, expression of D2S was increased (Po0.0001), expression levels between SCZ or BPD subjects with positive whereas expression levels of D2L (Po0.0001), D2Longer and negative toxicology tests. We also examined differences (Po0.0001) and D1 (Po0.0001) were decreased relative to between controls and SCZ patients positive and negative for controls (Figure 1). Moreover, there was a greater ratio of D2S to APDs, and found that for all transcripts examined patient groups D2L in the DLPFC of patients with SCZ (Po0.0001). positive and negative for APDs show significantly different In sharp contrast to SCZ, BPD patients showed an upregulation expression than controls. In particular, D2S, D2L, D2Longer and of D2L (Po0.0001) and DRD1 (Po0.0001). In patients with D1 expression levels were significantly different between SCZ with MDD expression of D2S (Po0.0001) was decreased as compared negative APD toxicology results vs controls (all P-values for Tukey with controls. Also, the ratio of D2S to D2L was smaller in patients honest significant difference o0.05), as well as between SCZ with BPD (Po0.0001) and MDD (Po0.0001) than controls. patients with positive APD toxicology results vs controls (all Comparisons between patient groups showed a relatively P-values for Tukey honest significant difference o0.05). higher expression of D2S in patients with SCZ when compared to patients with BPD (Po0.0001) and patients with MDD (Po0.0001). Furthermore, expression of D2L, D2Longer and DRD2 variants and DRD1 expression patterns during life span DRD1 were lower in patients with SCZ when compared to patients SCZ is thought to be a neurodevelopmental disorder character- with BPD and MDD (all Po0.0001). The ratio of D2S to D2L was ized by abnormalities in the generation and maturation of

DLPFC Hippocampus Caudate 3 *** *** 12 024 0123 DRD1 log2(DRD1) log2(DRD1) log2(DRD1) −2 −2 −1 −3 −3 −2 −1 0 −4

CTR BPD MDD SCZ CTR BPD MDD SCZ CTR BPD SCZ * 3 *** *** 2 1 0 log2(D2S) log2(D2S) log2(D2S) −2 −1 −3 −2 −1 0 1 2 3 −3 −4 −2 0 2 4

CTR BPD MDD SCZ CTR BPD MDD SCZ CTR BPD SCZ ** 3 *** *** 12 log2(D2L) log2(D2L) log2(D2L) −2 0 2 4 −3 −2 −1 0 1 2 3 −3 −2 −1 0 −4

CTR BPD MDD SCZ CTR BPD MDD SCZ CTR BPD SCZ * *** 24 D2Longer D2L D2S log2(D2Longer) log2(D2Longer) −3−2−10123 −4 −2 0

CTR BPD MDD SCZ CTR BPD MDD SCZ

Figure 1. Diagnostic differences in the expression of dopamine receptor 1 (DRD1) and dopamine receptor 2 (DRD2) transcripts. Mean mRNA expression of DRD1, D2S, D2-long (D2L), and D2Longer (depicted in rows), in the dorsolateral prefrontal cortex (DLPFC), hippocampus and caudate nucleus (depicted in columns), respectively. The data are expressed as log2 of expression levels normalized to the geometric mean of three housekeeping genes. CTR, controls; BPD, bipolar disorder; MDD, major depressive disorder; SCZ, schizophrenia. The asterisks over the bars indicate signiﬁcant differences as compared with controls: *Po0.05, **Po0.001, ***Po0.0001.

functional circuits including dopamine signaling. Thus, given regions. Elucidating developmental characteristics of dopa- regional and transcript specificity of differences in DRD1 and DRD2 mine receptor, D1 and D2 transcripts may shed light on expression between the diagnostic groups, we investigated diagnosis-specific changes. whether there are differences in developmental trajectories We found differential expression profiles for DRD1 and DRD2 between the DLPFC and the HPC, and between DRD1 and DRD2 receptors between the DLPFC and HPC. In the DLPFC, expression transcripts examined in this study. Such differences would suggest levels of D2S (F(1,319) = 11.47, Po0.001) and D2L (F(1,320) = different rates of maturation between transcripts and brain 24.92, Po0.0001) were significantly and negatively associated

Figure 2. Life span expression patterns of dopamine receptor 2 (DRD2) variants and dopamine receptor 1 (DRD1). Developmental expression patterns of DRD2 and DRD1 transcripts in the postmortem dorsolateral prefrontal cortex (DLPFC) and the hippocampus (HPC) of the normal controls. Expression of D1, D2S, D2-long (D2L), and D2Longer are shown as a function of age ranging from 14th to 20th fetal week (left part of the x axis depicted in weeks), and from birth to 85 years of age (right part of the x axis depicted in years). Each dot represents one subject. The gray lines show LOESS ﬁt across fetal and postnatal life span, respectively.

Molecular Psychiatry (2014), 1258 – 1266 © 2014 Macmillan Publishers Limited DRD1 and DRD2 expression in psychiatric disorders SS Kaalund et al 1263 with age. The expression of these transcripts was significantly reported for the D1 receptor binding in an imaging study by lower during the fetal period relative to postnatal life (D2S:F Rieckmann et al.26 (4,319) = 9.85, Po0.0001; fetal vs 0–20 years, Po0.0001; fetal vs >20 years Po0.001) and D2L (F(4,320) = 13.32, Po0.0001; fetal vs 0–20 years, Po0.0001; fetal vs >20 years, Po0.0001), peaked Association of genetic variants of DRD2 and DRD1 with expression shortly after birth, and then decreased significantly with age (D2S: Several common genetic variants of the DRD2 and DRD1 genes beta regression coefficient = − 0.016, Po0.05; D2L: beta regression have previously been associated with increased risk for SCZ and coefficient = − 0.014, Po0.05; see Figure 2). Expression pattern of cognitive deficits.8,10,19 We analyzed associations of expression of D2Longer resembled that of D2S and D2L, but did not correlate DRD2 splice variants with 147 SNPs within a region encompassing significantly with age. DRD1 expression increased dramatically DRD2 and DRD1 mRNA with 131 SNPs (Cis SNPs defined as 100 kb during fetal life (beta regression coefficient = 14.25, Po0.0001), up- and downstream of the transcription start site for both genes), and then remained stable from birth until after the age of 20 years including previously described SNPs associated with increased risk (P>0.05); thereafter DRD1 expression declined with age (beta for SCZ and/or cognitive deficits (Supplementary Table S3). The regression coefficient = − 0.01, Po0.0001). Overall expression analyses were performed only in control subjects in order to limit levels were higher in the two postnatal stages (F(2,312) = 40.39, the number of confounding factors (such as medication or other Po0.0001; fetal vs 0–20 years, Po0.0001, fetal vs >20 years, illness-related parameters). Although the gene-wide analysis did Po0.0001), and lower in individuals above 20 years of age not show significant associations between SNPs and expression compared with individuals 0–20 years of age (Po0.05). In (P corrected for multiple comparisons), the results from the a priori contrast, there were no significant age-related changes in analysis of risk and/or expression-associated DRD2 SNPs con- expression of DRD2 splice variants and DRD1 in the HPC (from firmed and extended previous results (Table 3). the fetal period to old age) and caudate nucleus (examined only in SNPs rs2283265 and rs1076560 are in complete linkage adulthood). disequilibrium with the SCZ-associated risk SNP rs1079727,8 and Finally, there were no significant interregional correlations of with rs6278, which has not been examined before. In agreement transcript expression (all r-values o0.2, all P-values >0.05) as was with the previous studies,10 carriers of the major alleles of rs1076560 and rs2283265 (C and G, respectively) showed higher expression of D2S relative to D2L as compared with minor allele carriers (A and T, respectively, Po0.05; see Figures 3 and 4). Furthermore, major allele carriers at rs2283265 and rs6278 (A and Table 3. Associations of DRD2 SNPs with expression of DRD2 variants G, respectively) showed higher expression of D2S relative to D2L as (a priori analyses) compared with minor allele carriers (G and T; Po0.05 and Po0.05, respectively). Thus, the expression changes in D2S/D2L SNP name DRD2 splice variant (P-values for SNP-expression observed in control subjects carrying SCZ risk-associated alleles associations) are consistent with the direction of changes in patients with SCZ in our study. D2S/D2L D2S D2L D2Longer Carriers of the rs2242592 risk allele (C),27 showed lower expression of D2Longer than major allele (T) carriers (Po0.05). rs6277 NS NS NS NS Similarly, carriers of the minor allele (T) of rs6275, which is rs1801028 NS NS NS NS 28 o considered a risk allele, showed higher expression of D2Longer rs6275 NS NS NS 0.05 than carriers of the major allele (C) (Po0.05). These results are rs1799732 NS NS NS NS fi rs2242592 NS NS NS o0.05 again consistent with the signi cantly lower expression of rs1079727 o0.05 NS NS NS D2Longer in patients with SCZ as compared with controls. rs2283265 o0.05 NS NS NS We did not confirm the association of risk alleles of rs6277 and rs12364283 NS NS NS NS rs179973229,30 with increased expression of DRD2. Moreover, there rs1076560 o0.05 NS NS NS were no significant associations (either corrected for multiple rs6278 o0.05 NS NS NS comparisons or nominally significant) between any SNPs exam- Abbreviations: D2L, D2-long; DRD2, dopamine 2 receptor; NS, not ined in this study and mRNA expression levels of DRD1 and DRD2 significant; SNP, single-nucleotide polymorphisms. P>0.05. splice variants in the HPC and caudate in all subjects or in any of the diagnostic or race group separately.

RISK SNPs rs6277

rs6275

rs2242592 rs1801028 rs2283265 rs1079727 rs1799732

8 7 6’ 6 5 4 3 2 1

rs1799732 Expression associated SNPs rs1076560 rs2283265 rs1079727 rs12364283 Figure 3. Dopamine receptor 2 (DRD2) single-nucleotide polymorphisms (SNPs). A schematic representation of the DRD2 gene and locations of SNPs. Coding exons are shown as gray boxes and noncoding exons as white boxes. The approximate locations of SNPs associated with increased risk for schizophrenia are shown above the gene, and the SNPs that have been associated with expression of DRD2 are shown below the gene.

© 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 1258 – 1266 DRD1 and DRD2 expression in psychiatric disorders SS Kaalund et al 1264 late puberty, suggesting that this receptor has a pivotal role during early postnatal brain development when maturation of DRD1 expression parallels the development of some cognitive features and coincides with the onset of psychiatric diseases. It is possible that lower expression levels of the postsynaptic receptors D2L and DRD1 reduce the modulatory role of dopamine in the PFC. Further, the increase in D2S may represent a dysregulation of the cortical output to the striatum, as presynaptic D2 receptors on corticostriatal afferents are thought to modulate the prefrontal excitatory neurotransmission.32–36 Prefrontal cortical glutamatergic axon-terminals in the striatum are positive for D2 receptor staining, and binding of dopamine to the D2 postsynaptic receptor decreases the release of glutamate.32 The observed increase in D2S expression in DLPFC could potentially reflect an increase in the D2 receptors on the glutamatergic postsynapse, which may be a compensatory mechanism to the elevated dopamine levels in striatum. However, these interpreta- tions are speculative, as we have only examined these receptors at the transcriptional level. The identities of D2S and D2L as presynaptic and postsynaptic receptors, respectively, are still somewhat unclear. The attribution of receptor functions are derived mainly from genetically modified mice. Knockout studies of D2L in mice have shown that this receptor is responsible for many of the Figure 4. Expression of ratio of D2S/D2L in normal control subjects postsynaptic functions of dopamine receptors in the striatum, as a function of genotype at dopamine receptor 2 (DRD2) schizo- whereas presynaptic responses were preserved, suggesting that 12,17,37 phrenia-associated risk single-nucleotide polymorphisms (SNPs). The these are harbored by the short D2 isoform. An expression expression ratio of D2S/D2L in the dorsolateral prefrontal cortex study in primates also favours a role for D2S as a presynaptic (DLPFC) is shown as a function of rs6278 (G/T), rs10786560 (A/C), receptor in both dopaminergic and non-dopaminergic neurons.16 rs2283265 (G/T), and rs1079727 (A/G) genotype. These SNPs are in Furthermore, the interaction between the D2 receptor and perfect linkage disequilibrium (LD). Each square represents one presynaptic markers such as DAT and tyrosine hydroxylase fi subject. The asterisk depicts a signi cant difference between the phosphorylation may be specific to D2S,17,38 and interaction groups (Po0.05). The horizontal bar shows the mean mRNA β expression of the group ± standard error. between the D2 receptor, DARP-32, -arrestin and PP-2A may be specific to D2L.17,39,40 Studies in cells overexpressing either D2S or D2L receptors that show their differential regulation by dopamine may elucidate why expression of these receptors is differentially altered in patients with SCZ. Pretreatment with dopamine upregulated the expres- DISCUSSION sion of D2L mRNA and downregulated D2S mRNA.41 These data In this study, we show that expression of the presynaptic D2 suggest that the low D2L levels observed in SCZ in our study may receptor, D2S, was increased, whereas the expression of long be a consequence of hypodopaminergic tone in the DLPFC.42 The variants, D2L, D2Longer as well as DRD1, which are predominantly development of isoform-specific antibodies for use in human postsynaptic, were decreased in the DLPFC of patients with SCZ tissue will be crucial for the evaluation of the possible significance compared with controls. The changes were brain region specific, of the splice variant-specific changes at the protein level. suggestive of predominantly prefrontal abnormalities in D1 and Another important issue to address is the possible confounding D2 receptors mRNA expression in SCZ. Perturbations of dopamine effects of medication. In keeping with previous postmortem signaling in the prefrontal cortex, in particular of D1 receptor studies,43–47 we did not observe an effect of APDs on the function31 but also DRD2 splicing,10 can contribute to cognitive expression of DRD2 and DRD1 in DLPFC of patients with SCZ or deficits, a core symptom of SCZ. This notion is further BPD. In experimental settings, APDs in the brains of rodents and corroborated by the genetic findings, showing that the SCZ risk primates increase the levels of D2 receptors both on mRNA and alleles of rs1079727, rs1076560 and rs2283265 were associated protein level.48–50 These studies did not examine effects of APDs with the relatively higher expression of D2S to D2L in the DLPFC of on DRD2 splicing, however. Whether D2S and D2L interact nonpsychiatric control subjects. We show that the directionality of differently with APDs is debatable. Some studies do not find the change in the expression ratio in risk allele carriers is differences in binding affinities for the two isoforms,51 whereas, consistent with the change observed in patients with SCZ, others do.52,53 Malmberg et al. 52 found that D2S had a higher suggesting a mechanism by which genetic variants in DRD2 may affinity for several APDs including clozapine in vitro (in transfected confer risk of developing SCZ. Furthermore, we show that the mice fibroblast cell line and rat pituitary cell line). In contrast, Xu expression changes of DRD2 variants and DRD1 in DLPFC were et al. 53 found that D2L had higher affinity for clozapine and specific to patients with SCZ, as patients with affective disorders haloperidol than D2S when measured in striatal tissue homo- showed changes in the opposite direction. These results imply genates. In the same study, clozapine was equally potent in that even though the patient groups share clinical symptoms, and D2L− / − and wild-type mice (90% of the striatal D2 receptors are to some extent medication treatment, different molecular D2L in this mice strain), suggesting that other brain areas than mechanisms underlie these disorders. striatum are critical for the APD effects of clozapine. Interestingly, We also found that life span expression of DRD2 and DRD1 Lidow et al. 50 show that the effect of clozapine (and olanzapine) transcripts was age- and region specific, possibly reflecting specifically increases the DRD2 mRNA levels in PFC but not in differences in the temporal development patterns of DLPFC, striatum in nonhuman primates. The effect was the same for the HPC and caudate nucleus. Of special interest is the expression two splice variants. Curiously, the non-APD D2-antagonist tiapiride pattern of DRD1 in the DLPFC, which increases dramatically during also increased the level of D2L in the PFC, whereas D2S mRNA prenatal brain development and reaches adult expression levels in levels were unaffected, perhaps suggesting an interaction

Molecular Psychiatry (2014), 1258 – 1266 © 2014 Macmillan Publishers Limited DRD1 and DRD2 expression in psychiatric disorders SS Kaalund et al 1265 between APDs and prefrontal D2S levels. Only a single study has 4 Kumakura Y, Cumming P, Vernaleken I, Buchholz HG, Siessmeier T, Heinz A et al. addressed the issue of APD effect on DRD1 on expression, and Elevated [18F]fluorodopamine turnover in brain of patients with schizophrenia: found that D1R is downregulated after chronic APD exposure in an [18F]fluorodopa/positron emission tomography study. J Neurosci 2007; 27: the nonhuman primate cortex.49 Human imaging data are 8080–8087. consistent with this finding.54 5 Hietala J, Syvälahti E, Vuorio K, Räkköläinen V, Bergman J, Haaparanta M et al. The relevance of altered splicing as a pathological mechanism Presynaptic dopamine function in striatum of neuroleptic-naive schizophrenic patients. Lancet 1995; 346: 1130–1131. for SCZ has previously been demonstrated by us and others. There 6 Abi-Dargham A, Rodenhiser J, Printz D, Zea-Ponce Y, Gil R, Kegeles LS et al. have been at least seven recent reports of differential expression Increased baseline occupancy of D2 receptors by dopamine in schizophrenia. Proc of splice variants of SCZ susceptibility genes in postmortem brain Natl Acad Sci USA 2000; 97: 8104–8109. tissues obtained from patients with SCZ, including DARPP32 7 Wong DF, Wagner HN, Tune LE, Dannals RF, Pearlson GD, Links JM et al. Positron (PPP1R1B),24 DISC155, ERBB456, GRM357 and SLC12A5 (KCC2).58 More- emission tomography reveals elevated D2 dopamine receptors in drug-naive over, another report using high-throughput RNA sequencing to schizophrenics. Science 1986; 234: 1558–1563. identify alternatively spliced genes in brain tissue in SCZ found 8 Glatt SJ, Faraone SV, Lasky-Su JA, Kanazawa T, Hwu HG, Tsuang MT. Family-based over a thousand genes, which were alternatively spliced in the association testing strongly implicates DRD2 as a risk gene for schizophrenia in 14 – superior temporal gyrus,59 including previously reported CAMK2, Han Chinese from Taiwan. Mol Psychiatry 2009; :885 893. 9 Glatt SJ, Faraone SV, Tsuang MT. Meta-analysis identifies an association between DISC1, GRIN (NR1), ERBB4 and FYN. the dopamine D2 receptor gene and schizophrenia. Mol Psychiatry 2003; 8:911–915. In summary, we found that expression levels of D2S, D2L, 10 Zhang Y, Bertolino A, Fazio L, Blasi G, Rampino A, Romano R et al. Polymorphisms in D2Longer and DRD1 transcripts are changed in patients with SCZ human dopamine D2 receptor gene affect gene expression, splicing, and neuronal and affective disorders in a diagnosis- and region-specific manner. activity during working memory. Proc Natl Acad Sci USA 2007; 104: 20552–20557. The expression of DRD2 transcripts was associated with genotypes 11 Bertolino A, Taurisano P, Pisciotta NM, Blasi G, Fazio L, Romano R et al. Genetically predicting cognitive function and/or risk of SCZ, with risk alleles determined measures of striatal D2 signaling predict prefrontal activity during showing changes in the same direction as that observed in SCZ. It working memory performance. PLoS ONE 2010; 5: e9348. is important to recognize the complexity of SNP-expression 12 Usiello A, Baik JH, Rougé-Pont F, Picetti R, Dierich A, LeMeur M et al. Distinct 408 relationships, and note that many other factors may influence functions of the two isoforms of dopamine D2 receptors. Nature 2000; : – these relationships, including epistatic effects, a large number of 199 203. 13 Seeman P, Nam D, Ulpian C, Liu IS, Tallerico T. New dopamine receptor, D2 environmental events as well as developmental factors. Thus, not (Longer), with unique TG splice site, in human brain. Brain Res Mol Brain Res 2000; detecting relationships between SNPs and gene expression levels 76:132–141. does not necessarily preclude the role that these SNPs have at a 14 Xie J, Lee JA, Kress TL, Mowry KL, Black DL. Protein kinase A phosphorylation certain time in the individual’s life. Another limitation of our study modulates transport of the polypyrimidine tract-binding protein. Proc Natl Acad is the use of tissue homogenates where cell type-specific changes Sci USA 2003; 100: 8776–8781. may be masked if they occur in low abundant cell types or if the 15 Sasabe T, Futai E, Ishiura S. Polypyrimidine tract-binding protein 1 regulates the changes are opposite between individual cell types. Such alternative splicing of dopamine receptor D(2). J Neurochem 2011; 116:76–81. concerns are especially relevant with regard to the interpretation 16 Khan ZU, Mrzljak L, Gutierrez A, de la Calle A, Goldman-Rakic PS. Prominence of fi the dopamine D2 short isoform in dopaminergic pathways. Proc Natl Acad Sci USA of negative ndings, for example, of DRD2 variants and expression 95 – in HPC and caudate. Nevertheless, the results reflect on the 1998; :7731 7736. 17 Lindgren N, Usiello A, Goiny M, Haycock J, Erbs E, Greengard P et al. Distinct roles complexity of dopamine receptor splicing, and raise the possibility of dopamine D2L and D2S receptor isoforms in the regulation of protein phos- that the relative abundance of splice variants, not only total phorylation at presynaptic and postsynaptic sites. Proc Natl Acad Sci USA 2003; amounts of D2 receptors, are relevant to the pathophysiology of 100: 4305–4309. SCZ and affective disorders. 18 Takahashi H, Yamada M, Suhara T. Functional significance of central D1 receptors in cognition: beyond working memory. J Cereb Blood Flow Metab 2012; 32:1248–1258. 19 Zhu F, Yan CX, Wang Q, Zhu YS, Zhao Y, Huang J et al. An association study CONFLICT OF INTEREST between dopamine D1 receptor gene polymorphisms and the risk of schizo- 1420 – The authors declare no conflict of interest. phrenia. Brain Res 2011; :106 113. 20 Rothmond DA, Weickert CS, Webster MJ. Developmental changes in human dopamine neurotransmission: cortical receptors and terminators. BMC Neurosci ACKNOWLEDGMENTS 2012; 13:18. 21 Nieoullon A. Dopamine and the regulation of cognition and attention. Prog We thank Liqin Wang and Vesna Imamovic for their technical expertise, and Neurobiol 2002; 67:53–83. fi Dr Llewellyn Bigelow for his diagnostic contributions. We thank the Of ces of the 22 Morikawa T, Manabe T. Aberrant regulation of alternative pre-mRNA splicing in Chief Medical Examiner of Washington, DC and of Northern Virginia-Northern District, schizophrenia. Neurochem Int 2010; 57: 691–704. and Dr Ronald Zielke, Robert Johnson and Robert Vigorito at the National Institute of 23 Lipska BK, Deep-Soboslay A, Weickert CS, Hyde TM, Martin CE, Herman MM et al. Child Health and Human Brain Development Brain and Tissue Bank for Develop- Critical factors in gene expression in postmortem human brain: Focus on studies mental Disorders, University of Maryland School of Medicine, as well as Dr Maree in schizophrenia. Biol Psychiatry 2006; 60: 650–658. Webster at the Stanley Medical Institute for their collection of brains. We also thank 24 Kunii Y, Hyde TM, Ye T, Li C, Kolachana B, Dickinson D et al. Revisiting the families of the deceased for the donations of brain tissue and their time and DARPP-32 in postmortem human brain: changes in schizophrenia and bipolar effort devoted to the consent process and interviews. This research was supported by disorder and genetic associations with t-DARPP-32 expression. Mol Psychiatry the Intramural Research Program of the National Institute of Mental Health at the advance online publication, 8 January 2013; doi: 10.1038/mp.2012.174 (e-pub National Institutes of Health, and by the Copenhagen Graduate School of Health ahead of print). Science, University of Copenhagen and Julie von Müellens Foundation (to SSK). 25 Colantuoni C, Lipska BK, Ye T, Hyde TM, Tao R, Leek JT et al. Temporal dynamics and genetic control of transcription in the human prefrontal cortex. Nature 2011; 478:519–523. REFERENCES 26 Rieckmann A, Karlsson S, Karlsson P, Brehmer Y, Fischer H, Farde L et al. Dopamine 1 Creese I, Burt DR, Snyder SH. Dopamine receptor binding predicts clinical and D1 receptor associations within and between dopaminergic pathways in younger pharmacological potencies of antischizophrenic drugs. Science 1976; 192: and elderly adults: links to cognitive performance. Cereb Cortex 2011; 21: 481–483. 2023–2032. 2 Seeman P, Lee T. Antipsychotic drugs: direct correlation between clinical potency 27 Dubertret C, Bardel C, Ramoz N, Martin PM, Deybach JC, Adès J et al. A genetic and presynaptic action on dopamine neurons. Science 1975; 188: 1217–1219. schizophrenia-susceptibility region located between the ANKK1 and DRD2 genes. 3 Laruelle M, Abi-Dargham A, van Dyck CH, Gil R, D'Souza CD, Erdos J et al. Single Prog Neuropsychopharmacol Biol Psychiatry 2010; 34:492–499. photon emission computerized tomography imaging of amphetamine-induced 28 Allen NC, Bagade S, McQueen MB, Ioannidis JP, Kavvoura FK, Khoury MJ et al. dopamine release in drug-free schizophrenic subjects. Proc Natl Acad Sci USA Systematic meta-analyses and field synopsis of genetic association studies in 1996; 93: 9235–9240. schizophrenia: the SzGene database. Nat Genet 2008; 40:827–834.

© 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 1258 – 1266 DRD1 and DRD2 expression in psychiatric disorders SS Kaalund et al 1266 29 Arinami T, Gao M, Hamaguchi H, Toru M. A functional polymorphism in the schizophrenia: effect of antipsychotic treatment. Psychopharmacology 2009; 206: promoter region of the dopamine D2 receptor gene is associated with schizo- 313–324. phrenia. Hum Mol Genet 1997; 6: 577–582. 46 Zhan L, Kerr JR, Lafuente MJ, Maclean A, Chibalina MV, Liu B et al. Altered 30 Duan J, Wainwright MS, Comeron JM, Saitou N, Sanders AR, Gelernter J et al. expression and coregulation of dopamine signalling genes in schizophrenia and Synonymous mutations in the human dopamine receptor D2 (DRD2) affect mRNA bipolar disorder. Neuropathol Appl Neurobiol 2011; 37: 206–219. stability and synthesis of the receptor. Hum Mol Genet 2003; 12:205–216. 47 Rothmond D, Akil M, Weickert C. Alterations in Dopamine d2 receptor isoform 31 Sawaguchi T, Goldman-Rakic PS. D1 dopamine receptors in prefrontal cortex: gene expression in the DLPFC of patients with schizophrenia. Annual Meeting for involvement in working memory. Science 1991; 251:947–950. Society for Neuroscience 2010 40th Annual Meeting: Poster 326.316/Z313. 32 Bamford NS, Zhang H, Schmitz Y, Wu NP, Cepeda C, Levine MS et al. Hetero- 48 Janowsky A, Neve KA, Kinzie JM, Taylor B, de Paulis T, Belknap JK. Extrastriatal synaptic dopamine neurotransmission selects sets of corticostriatal terminals. dopamine D2 receptors: distribution, pharmacological characterization and Neuron 2004; 42: 653–663. region-specific regulation by clozapine. J Pharmacol Exp Ther 1992; 261: 33 Centonze D, Gubellini P, Usiello A, Rossi S, Tscherter A, Bracci E et al. Differential 1282–1290. contribution of dopamine D2S and D2L receptors in the modulation of glutamate 49 Lidow MS, Elsworth JD, Goldman-Rakic PS. Down-regulation of the D1 and D5 and GABA transmission in the striatum. Neuroscience 2004; 129:157–166. dopamine receptors in the primate prefrontal cortex by chronic treatment with 34 Wang JQ, McGinty JF. Differential effects of D1 and D2 dopamine receptor antipsychotic drugs. J Pharmacol Exp Ther 1997; 281:597–603. antagonists on acute amphetamine- or methamphetamine-induced up-regula- 50 Lidow MS, Goldman-Rakic PS. Differential regulation of D2 and D4 dopamine tion of zif/268 mRNA expression in rat forebrain. J Neurochem 1995; 65: receptor mRNAs in the primate cerebral cortex vs neostriatum: effects of chronic 2706–2715. treatment with typical and atypical antipsychotic drugs. J Pharmacol Exp Ther 35 Brady AM, O'Donnell P. Dopaminergic modulation of prefrontal cortical input to 1997; 283:939–946. nucleus accumbens neurons in vivo. J Neurosci 2004; 24: 1040–1049. 51 Leysen JE, Gommeren W, Mertens J, Luyten WH, Pauwels PJ, Ewert M et al. 36 O'Donnell P, Grace AA. Tonic D2-mediated attenuation of cortical excitation in Comparison of in vitro binding properties of a series of dopamine antagonists and nucleus accumbens neurons recorded in vitro. Brain Res 1994; 634: 105–112. agonists for cloned human dopamine D2S and D2L receptors and for D2 37 De Mei C, Ramos M, Iitaka C, Borrelli E. Getting specialized: presynaptic and receptors in rat striatal and mesolimbic tissues, using [125I] 2'-iodospiperone. postsynaptic dopamine D2 receptors. Curr Opin Pharmacol 2009; 9:53–58. Psychopharmacology (Berl) 1993; 110:27–36. 38 Lee FJ, Pei L, Moszczynska A, Vukusic B, Fletcher PJ, Liu F. Dopamine transporter 52 Malmberg A, Jackson DM, Eriksson A, Mohell N. Unique binding characteristics of cell surface localization facilitated by a direct interaction with the dopamine D2 antipsychotic agents interacting with human dopamine D2A, D2B, and D3 receptor. EMBO J 2007; 26: 2127–2136. receptors. Mol Pharmacol 1993; 43: 749–754. 39 Beaulieu JM, Sotnikova TD, Marion S, Lefkowitz RJ, Gainetdinov RR, Caron MG. An 53 Xu R, Hranilovic D, Fetsko LA, Bucan M, Wang Y. Dopamine D2S and D2L receptors Akt/beta-arrestin 2/PP2A signaling complex mediates dopaminergic neuro- may differentially contribute to the actions of antipsychotic and psychotic agents transmission and behavior. Cell 2005; 122:261–273. in mice. Mol Psychiatry 2002; 7:1075–1082. 40 Beaulieu JM, Tirotta E, Sotnikova TD, Masri B, Salahpour A, Gainetdinov RR et al. 54 Abi-Dargham A, Xu X, Thompson JL, Gil R, Kegeles LS, Urban N et al. Increased Regulation of Akt signaling by D2 and D3 dopamine receptors in vivo. J Neurosci prefrontal cortical D(1) receptors in drug naive patients with schizophrenia: a PET 2007; 27: 881–885. study with [(1)(1)C]NNC112. J Psychopharmacol 2012; 26: 794–805. 41 Zhang LJ, Lachowicz JE, Sibley DR. The D2S and D2L dopamine receptor isoforms 55 Nakata K, Lipska BK, Hyde TM, Ye T, Newburn EN, Morita Y et al. DISC1 splice are differentially regulated in Chinese hamster ovary cells. Mol Pharmacol 1994; variants are upregulated in schizophrenia and associated with risk polymorph- 45: 878–889. isms. Proc Natl Acad Sci USA 2009; 106: 15873–15878. 42 Akil M, Pierri JN, Whitehead RE, Edgar CL, Mohila C, Sampson AR et al. Lamina- 56 Law AJ, Kleinman JE, Weinberger DR, Weickert CS. Disease-associated intronic specific alterations in the dopamine innervation of the prefrontal cortex in schi- variants in the ErbB4 gene are related to altered ErbB4 splice-variant expression in zophrenic subjects. Am J Psychiatry 1999; 156: 1580–1589. the brain in schizophrenia. Hum Mol Genet 2007; 16:129–141. 43 Roberts DA, Balderson D, Pickering-Brown SM, Deakin JF, Owen F. The abundance 57 Sartorius LJ, Weinberger DR, Hyde TM, Harrison PJ, Kleinman JE, Lipska BK. of mRNA for dopamine D2 receptor isoforms in brain tissue from controls and Expression of a GRM3 splice variant is increased in the dorsolateral prefrontal schizophrenics. Brain Res Mol Brain Res 1994; 25:173–175. cortex of individuals carrying a schizophrenia risk SNP. Neuropsychopharmacology 44 Meador-Woodruff JH, Haroutunian V, Powchik P, Davidson M, Davis KL, Watson SJ. 2008; 33: 2626–2634. Dopamine receptor transcript expression in striatum and prefrontal and occipital 58 Tao R, Li C, Newburn EN, Ye T, Lipska BK, Herman MM et al. Transcript-specific cortex. Focal abnormalities in orbitofrontal cortex in schizophrenia. Arch Gen associations of SLC12A5 (KCC2) in human prefrontal cortex with development, Psychiatry 1997; 54: 1089–1095. schizophrenia, and affective disorders. J Neurosci 2012; 32: 5216–5222. 45 Urigüen L, García-Fuster MJ, Callado LF, Morentin B, La Harpe R, Casadó V et al. 59 Wu JQ, Wang X, Beveridge NJ, Tooney PA, Scott RJ, Carr VJ et al. Transcriptome Immunodensity and mRNA expression of A2A adenosine, D2 dopamine, and sequencing revealed significant alteration of cortical promoter usage and splicing CB1 cannabinoid receptors in postmortem frontal cortex of subjects with in schizophrenia. PLoS ONE 2012; 7: e36351.

Supplementary Information accompanies the paper on the Molecular Psychiatry website (http://www.nature.com/mp)