Molecular and Genetic Evidence for Abnormalities in the Nodes of Ranvier in Schizophrenia

ORIGINAL ARTICLE

ONLINE FIRST Molecular and Genetic Evidence for Abnormalities in the Nodes of Ranvier in Schizophrenia

Panos Roussos, MD, PhD; Pavel Katsel, PhD; Kenneth L. Davis, MD; Panos Bitsios, MD, PhD; Stella G. Giakoumaki, PhD; Jigar Jogia, PhD; Kinga Rozsnyai, MSc; David Collier, PhD; Sophia Frangou, MD, PhD; Larry J. Siever, MD; Vahram Haroutunian, PhD

Context: Genetic, neuroimaging, and molecular neu- white men and women participated in the cognitive robiological evidence support the hypothesis that the dis- (n=513) and neuroimaging (n=52) studies. connectivity syndrome in schizophrenia (SZ) could arise from failures of saltatory conduction and abnormalities Main Outcome Measures: The mRNA and protein at the nodes of Ranvier (NOR) interface where myelin levels in postmortem brain samples, genetic association and axons interact. with schizophrenia, cognitive performance, and blood oxygenation level–dependent functional magnetic reso- Objective: To identify abnormalities in the expression nance imaging. of oligodendroglial genes and proteins that participate in the formation, maintenance, and integrity of the NOR Results: The mRNA expression of multiple NOR genes in SZ. was decreased in schizophrenia. The ANK3 rs9804190 C allele was associated with lower ANK3 mRNA expres- Design: The messenger RNA (mRNA) expression lev- sion levels, higher risk for SZ in the case-control cohort, els of multiple NOR genes were quantified in 2 indepen- and poorer working memory and executive function per- dent postmortem brain cohorts of individuals with SZ, formance and increased prefrontal activation during a and generalizability to protein expression was con- firmed. The effect of the ANK3 genotype on the mRNA working memory task in healthy individuals. expression level was tested in postmortem human brain. Case-control analysis tested the association of the ANK3 Conclusions: These results point to abnormalities in the genotype with SZ. The ANK3 genotype’s influence on cog- expression of genes and protein associated with the in- nitive task performance and functional magnetic reso- tegrity of the NOR and suggest them as substrates for the nance imaging activation was tested in 2 independent co- disconnectivity syndrome in SZ. The association of ANK3 horts of healthy individuals. with lower brain mRNA expression levels implicates a molecular mechanism for its genetic, clinical, and cog- Setting: Research hospital. nitive associations with SZ.

Patients: Postmortem samples from patients with SZ and Arch Gen Psychiatry. 2012;69(1):7-15. healthy controls were used for the brain expression study Published online September 5, 2011. (n=46) and the case-control analysis (n=272). Healthy doi:10.1001/archgenpsychiatry.2011.110

URING THE PAST DECADE, (NOR).4 The NOR span the entire length multiple lines of evi- of myelinated axons and comprise my- dence (gene and protein elin sheath–free segments where ion (so- expression, genetic asso- dium and potassium) exchange can take ciation, neuroimaging) place, propagating action potentials from haveD suggested that the disconnectivity the axon initial segment down the length syndrome thought to underlie the neuro- of the axon to synaptic terminals. In the biology of schizophrenia (SZ) is in part due central nervous system, NOR are formed to abnormalities in myelination and oli- and maintained through reciprocal inter- godendroglial function.1-3 Efficient and fast actions between neurons and oligoden- neurotransmission is dependent on salta- drocytes.5 Multiple and specific neuronal tory conduction, which is in turn depen- and oligodendroglial proteins such as an- Author Affiliations are listed at dent on the integrity of myelin sheaths, oli- kyrin G (AnkG) interact to ensure the high the end of this article. godendrocytes, and the nodes of Ranvier concentration (Ͼ1200 voltage-gated so-

ARCH GEN PSYCHIATRY/ VOL 69 (NO. 1), JAN 2012 WWW.ARCHGENPSYCHIATRY.COM 7

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 dium ion channels/µm2) and anchoring of voltage-gated val, tissue pH, and sex distributions of the subjects are shown sodium ion channels (Nav; Nav1.6 in adults) to the NOR in eTable 1 (http://www.archgenpsychiatry.com). and maintain tight junctions between the axolemma and Gray matter from 14 cortical-frontal (Brodmann area [BA] paranodal myelin loops.5 The AnkG protein belongs to 8, 10, 44, 46, 4), cingulate (BA 23/31, 24/32), parietal (BA 7), a family of scaffolding proteins that interact with sev- temporal (BA 20, 21, 22, 36/28, 38), and occipital (BA 17) brain regions and 3 noncortical brain regions (caudate, hippocam- eral proteins, including the cytoskeletal protein ␤IV spec- 5,6 pus, and putamen) were used for the Affymetrix HG-U133AB trin and Nav channels. The AnkG and Nav subunits GeneChip gene expression analysis (n=17-39) (Affymetrix, also interact in cis with neurofascin 186 and paranodal Santa Clara, California) as described in detail previously.7,8,24 contactin.5,6 Contactin and Caspr interact in trans with Similarly prepared aliquots from the superior temporal gyrus cell adhesion molecules of oligodendroglial paranodal (STG) and primary visual cortex were used in quantitative poly- loops that include the neurofascin 155 isoform and Tag1 merase chain reaction (qPCR) and quantitative Western blot- to ensure the integrity of the myelin-axolemma tight junc- ting analysis in a larger and independent set of brain tissue speci- tion.5,6 Abnormalities in the expression of oligodendrog- mens (eTable 1).25,26 lial and neuronal genes and proteins in SZ could ham- Additional studies for genetic association of ANK3 rs9804190 per saltatory conduction by affecting the formation, were carried out using case-control samples from the Brain Bank. For analytical purposes, only white American individuals with maintenance, and integrity of the NOR and contribute 4 Western European ancestry (SZ, n=208; comparison control, to the hypothesized disconnectivity syndrome. n=64) were used for the clinical genetic investigation of the The current studies were undertaken to advance the ANK3 rs9804190 polymorphism. observation of myelin and oligodendrocyte gene and protein expression abnormalities in SZ to the functional do- Healthy Individuals main and to the study of genetic association based on direct neurobiological observations in the brains of persons Learning on Genetics of Schizophrenia Spectrum Cohort and with SZ with specific gene and protein expression ab- Cognitive Task Performance. We examined the effect of the normalities. Exploratory analysis of data from a large post- ANK3 rs9804190 polymorphism using intermediates pheno- mortem microarray study of multiple brain regions of per- types in 530 individuals from the Learning On Genetics Of 7,8 Schizophrenia Spectrum (LOGOS) cohort. This cohort has been sons with SZ suggested the abnormal expression of many 27 neuron- and oligodendrocyte-enriched genes that en- described in detail previously. code proteins associated with paranodal junctions and 9 British Cohort and Neuroimaging Data. We further explored NOR. We therefore sought to identify the involvement the effect of the ANK3 rs9804190 polymorphism on brain ac- of axoglial interacting genes and proteins in SZ more di- tivity (on functional magnetic resonance imaging [fMRI]) dur- rectly by using postmortem brain tissue from an inde- ing a working memory task. Fifty-two healthy volunteers of white pendent cohort of persons with SZ and unaffected com- British descent were assessed with the Structured Clinical In- parison controls. terview for DSM-IV-TR Axis I Disorders28 and the Family In- ANK3 was among the genes that were abnormally ex- terview for Genetic Studies29 to exclude any personal psychi- pressed in independent gene and protein expression stud- atric history or family history (up to second-degree relatives) ies. Because recent genome-wide association studies have of SZ or affective disorders. Participants were further screened implicated ANK3 in bipolar disorder (BD)10-13 and SZ14 to exclude past, current, and hereditary medical disorders, and because several studies have suggested commonali- DSM-IV lifetime drug or alcohol dependence and drug or alcohol abuse in the preceding 6 months, and contraindications ties in the neurobiology,15-17 etiology, and genetic asso- 12,13,18,19 to MRI. For all participants, an estimate of the current full- ciations between SZ and BD, we also sought to de- scale IQ was obtained on the day of scanning using the Wechs- termine whether ANK3 genetic variants affected the ler Adult Intelligence Scale–Revised,30 while handedness was expression of ANK3 in the same SZ and control brain based on self-report. Written informed consent was obtained specimens and in a larger postmortem collection of ra- from all subjects before study participation. The study was ap- cially homogeneous cases and controls (n=272). The in- proved by the Joint Ethics Committee of the Institute of Psy- fluence of the SZ risk–associated ANK3 polymorphism chiatry and the South London and Maudsley NHS Foundation on cognitive domains of significance to SZ and on brain Trust. activity during performance of a working memory task was also examined in healthy controls (n=513 and n=52, LONG-TERM HALOPERIDOL STUDIES respectively). IN RATS

To assess the effects of neuroleptic exposure on the expres- METHODS sion of selected NOR genes of interest, groups of 8 male Sprague- Dawley rats (6-8 months of age) received daily subcutaneous SUBJECTS injections of haloperidol (2 mg/kg) or the saline vehicle for 21 days. Ank3, Nfasc, Nrcam, and Scn8a expressions were as- Postmortem Cohort sessed by the procedures described earlier using rat-specific prim- ers and probes (eTable 2). Brain tissue specimens were derived from the Brain Bank of the Department of Psychiatry, Mount Sinai School of Medicine, New MICROARRAY PROCEDURE York, New York/James J. Peters Veterans Affairs Medical Cen- AND DATA ANALYSIS ter, Bronx, New York. The precise tissue-handling procedures have been described in detail.20-23 The cause of death, type of Microarray analysis was performed using the Affymetrix HG- antipsychotic medications used, mean age, postmortem inter- U133AB GeneChip set as described previously.7,8 Statistical com-

ARCH GEN PSYCHIATRY/ VOL 69 (NO. 1), JAN 2012 WWW.ARCHGENPSYCHIATRY.COM 8

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 parisons were made using the GeneSpring GX 7.3.1 cross- THE N-BACK TASK IN NEUROIMAGING gene error model (Agilent Technologies/Silicon Genetics, Santa Clara), based on the deviation from 1.0 algorithm. Filtering con- The 3-back task was administered in a block design. Partici- ditions were a combination of confidence (PՅ.05) and fold pants were instructed to respond by button press to the target change (Ն1.6) with Benjamini and Hochberg31 multiple test- letter. In the baseline (0-back) condition, the designated tar- ing corrections. get letter was X. In the experimental condition (3-back), the target letter was defined as any letter that was identical to the Real-time qPCR one presented 3 trials back. In each condition, a series of 14 letters were visually presented for 2 seconds each. Responses The messenger RNA (mRNA) levels of selective NOR genes were were monitored via a magnetic resonance–compatible button measured by qPCR in a larger independent (from the micro- box held in the subject’s dominant hand. There were 12 ep- array cases and controls) cohort (eTable 1) using TaqMan probes ochs in all, each lasting 30 seconds, with a total experiment and primer sets (Applied Biosystems, Foster City, California). duration of 6 minutes. The ratio of target to nontarget letters TaqMan probe identification numbers for selected nodal genes presented per block ranged from 2:12 to 3:11. Reaction time and normalization controls are listed in eTable 2. For relative to target letters and number of correct responses (accuracy) were quantification of mRNA expression, geometric means of the ex- recorded. pression of 3 housekeeping genes were calculated using the stan- dard curve method. The housekeeping genes PGK1, PPIA, and IMAGE ACQUISITION RPLPO were selected for their stability after comparison of several different endogenous controls using geNorm (http://medgen Both anatomical and gradient-echo echoplanar imaging data were .ugent.be/~jvdesomp/genorm/). Two housekeeping genes (Gadph acquired during the same session using a 1.5-T neuro- and Ppia) were used as the endogenous references in the rat optimized Signa MRI system (General Electric, Milwaukee, Wis- studies. consin). A total of 150 T2*-weighted echoplanar imaging brain volumes depicting blood oxygenation level–dependent con- QUANTITATIVE WESTERN BLOTTING trast were acquired at each of the 16 axial planes (repetition time, 2000 milliseconds; echo time, 40 milliseconds; flip angle, Protein abundance was measured in the STG from patients with 70°; slice thickness, 7 mm; slice skip, 0.7 mm; matrix size, SZ and comparison control subjects (n=8/group) previously 64ϫ64; voxel dimensions, 3.75ϫ3.75ϫ7.7 mm). For the pur- selected for qPCR analysis using Western blotting. Blots were poses of anatomical localization, a T1-weighted 3-dimen- probed with mouse monoclonal anti-AnkG antibody (IgG1, 1:200 sional inversion recovery prepared spoiled gradient recalled se- vol/vol dilution; NeuroMab, Davis, California), which binds spe- quence was used to acquire structural images in 124 axial planes cifically to 2 different isoforms (approximately 270 kDa and (repetition time, 1800 milliseconds; echo time, 5.1 millisec- approximately 250 kDa) and rabbit anti–glyceraldehyde-3- onds; inversion time, 450 milliseconds; flip angle, 20°; slice thick- phosphate dehydrogenase antibody (1:5000 vol/vol dilution; ness, 1.5 mm; matrix size, 256ϫ192; field of view, 240ϫ180 LifeSpan Biosciences, Seattle, Washington). Visualization and mm; voxel dimensions, 0.9375ϫ0.9375ϫ1.5 mm; number of quantification of bands were performed with the Odyssey 2.1 excitations, 1). software (Li-COR Biosciences, Lincoln, Nebraska). To ac- count for gel-to-gel variability, the relative expression value of STATISTICAL ANALYSIS AnkG and glyceraldehyde-3-phosphate dehydrogenase in each sample was calculated as a ratio between the averaged inten- Multiple statistical procedures were used for different aspects sities of the band in the experimental sample and in the stan- of the study. Comparison of the demographic variables was per- dard calibrator (a mix of small aliquots of tissue from all formed using separate 1-way analyses of variance or the non- samples). parametric Mann-Whitney U test as appropriate, based on deviation from normality. A 2-tailed t test was used to compare DNA GENOTYPING AND CLINICAL the relative abundance of AnkG protein and the relative ANK3 ASSOCIATION ANALYSIS mRNA expression of NOR genes in qPCR and genotyping experiments using SPSS version 17 statistical software (SPSS Inc, Samples of DNA from the postmortem cohort were extracted Chicago, Illinois). from the STG using the Genomic DNA-Tissue MiniPrep kit We used QTPHASE software (http://www.mrc-bsu.cam.ac (Zymo Research, Irvine, California). We initially genotyped the .uk/personal/frank/software/unphased/) from the UNPHASED qPCR cohort (n=48) for ANK3 rs9804190, rs10994336, and package32 version 3.1.4 for the analysis of genotype associa- rs10761482 polymorphisms that contribute to SZ and BD tions in the LOGOS cohort and ANK3 expression levels in the risk10,11,14 (Polymorphic DNA Technologies, Alameda, Califor- qPCR cohort. The ANK3 rs9804190 polymorphism was exam- nia). Genotyping of the LOGOS cohort was performed blind ined for allelic association with SZ in the Brain Bank cohort to phenotype measures with a competitive allele-specific PCR with PLINK software33 (http://pngu.mgh.harvard.edu/~purcell system (KBioscience, Herts, England). Brain Bank cohort geno- /plink/). Quality control for the Brain Bank cohort was done types were analyzed by Children’s Hospital of Philadelphia, Phila- as described previously10; correction for population stratifica- delphia, Pennsylvania, using the Affymetrix single-nucleotide tion was achieved using the eigenstrat method.34 Logistic re- polymorphism 6.0 array. Genotyping of the British cohort was gression analysis was carried out to evaluate the association of determined by a TaqMan allelic discrimination assay (Applied rs9804190 with SZ. P values from the UNPHASED and PLINK Biosystems, assay C_2584015_10). Genotyping quality con- analyses were corrected for multiple testing by running 100 000 trol for ANK3 rs9804190 in both the Brain Bank and LOGOS permutations of the data. Additionally, to lower the possibil- cohorts was performed in more than 10% of the samples by du- ity of type I statistical error in the LOGOS analysis, we set the plicate checking (rate of concordance in duplicates Ͼ99%). Call level of significance to .01. rate was approximately 97% for all polymorphisms. The mi- For the fMRI data analysis, we used the Statistical Paramet- nor allele frequencies of the ANK3 rs9804190 variant were iden- ric Mapping (SPM5) software (http://www.fil.ion.ucl.ac.uk tical across the different cohorts (eTable 3). /spm/software/spm5) implemented in MATLAB 7.1 (MathWorks

ARCH GEN PSYCHIATRY/ VOL 69 (NO. 1), JAN 2012 WWW.ARCHGENPSYCHIATRY.COM 9

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 Table 1. Microarray-Based Gene Expression Analysis of Selected Nodal and Tight Junction Genesa

STG PVC

Gene Maximum P FC P FC Gene Ontology Sequence Cluster Symbol t Score Value Ratio Value Ratio Biological Process35 Upregulated in SZ Sodium channel, voltage-gated, type III, ␤ SCN3B 5.23 .66 −1.09 .18 1.28 Sodium ion transport Potassium voltage-gated channel, subfamily F, KCNF1 5.14 .02b 1.32 .79 1.04 Cation transport member 1 Sodium channel, voltage-gated, type III, ␤ SCN3B 4.94 .42 −1.15 .24 1.30 Sodium ion transport Potassium voltage-gated channel, KQT-like KCNQ2 4.74 Ͼ.99 −1.00 .18 1.22 Cation transport subfamily Potassium voltage-gated channel, ␤ member 1/2 KCNAB1 4.50 .23 −1.18 .97 −1.01 Cation transport Contactin 1 CNTN1 4.21 .51 −1.12 .22 1.23 Notch signaling pathway, cell adhesion Potassium voltage-gated channel, subfamily H, KCNH7 3.71 .01b 1.36 .45 1.13 Cation transport member 7 Sodium channel, voltage-gated, type X, ␣ SCN10A 3.36 Ͻ.01b 1.74 .51 −1.13 Cation transport Contactin-associated protein-like 2, CASPR2 CNTNAP2 3.06 .01b 1.45 .31 1.17 Cell adhesion Sodium channel, voltage-gated, type II, ␤ SCN2B 1.85 .93 −1.01 .83 1.05 Sodium ion transport Contactin-associated protein 1, CASPR1 CNTNAP1 1.79 .68 −1.04 .49 −1.09 Cell adhesion Sodium channel, voltage-gated, type III, ␣ SCN3A 1.73 .05 −1.34 .53 −1.11 Sodium ion transport Downregulated in SZ Neurofascin NFASC −8.12 .04b −1.26 .27 −1.10 Cell adhesion Claudin 11 CLDN11 −7.29 .009b −1.88 .18 −1.61 Calcium-independent cell-cell adhesion Myelin-associated glycoprotein MAG −7.24 .25 −1.28 .93 −1.03 Cell adhesion Tight junction protein 2 TJP2 −6.35 .02b −1.54 .15 −1.39 Intercellular junction assembly Contactin 2 CNTN2, −6.10 .52 −1.13 .49 −1.11 Cell adhesion TAG1 Ankyrin G ANK3 −5.43 Ͻ.001b −1.52 .23 1.16 Cytoskeletal anchoring Contactin 3 CNTN3 −2.53 .007b −1.41 .13 −1.33 Cell adhesion Neuronal cell adhesion molecule NRCAM −2.52 .31 −1.13 .61 1.07 Cell-cell adhesion Tight junction protein 1 TJP1 −2.09 .24 −1.13 .53 −1.08 Intercellular junction assembly

Abbreviations: FC, fold change; PVC, primary visual cortex; STG, superior temporal gyrus; SZ, schizophrenia. a Data represent aggregate expression change scores across 17 different brain regions. Sample sizes vary for the different brain regions analyzed. The maximum t score is an extension of the fold change algorithm that was used as a standardized measure of gene expression change in SZ across all analyzed brain regions.7 Analysis of variance P Ͻ .01 for all of the listed genes. Upregulated and downregulated genes are listed in descending order of change score. The FC ratios and P values are from the same microarray collection separately for the STG and PVC. b P Ͻ.05.

Inc, Natick, Massachusetts) for data processing and analysis. the Marsbar toolbox (http://marsbar.sourceforge.net) to exam- Preprocessing of the images involved realignment to correct for ine the effect size of the genotype on the level of activation, ex- movement, normalization into Montreal Neurological Insti- pressed as Cohen d. tute space using the participant’s structural MRI images, and smoothing with an 8-mm full-width half-maximum isotropic gaussian kernel. A first-level fixed-effects model was com- RESULTS puted for each participant. The fMRI responses were con- volved with a canonical hemodynamic response function with GENE EXPRESSION OF NOR GENES the vectors of interest and a high-pass filter (128 seconds) was IS AFFECTED IN SZ applied to remove low-frequency noise. A hemodynamic response function–shaped low-pass filter was used to filter low- frequency noise artifacts. Contrast images representing the We initially conducted an exploratory analysis using mi- 3-back vs 0-back conditions were produced for each subject croarray data (contrast analysis t scores7) from 17 dif- and entered in the second-level, random-effects, 2-sample t test ferent brain regions representing all 4 cortical lobes, the analysis (group 1: CC; group 2: CT and TT). Results were con- hippocampus, and basal ganglia in persons with SZ (n=9- Ͻ sidered significant at P .05. False discovery rate corrected for 21) vs unaffected comparison controls (n=8-18) (eTable multiple comparisons. Stereotactic coordinates of the peak 1). We found significant dysregulation in the expres- maxima of the suprathreshold clusters were converted from the sion of genes known to participate in the development, Montreal Neurological Institute spatial array to that of Ta- lairach and Tournoux (http://imaging.mrc-cbu.cam.ac.uk organization, and maintenance of the NOR across all brain /imaging/MniTalairach), and corresponding BAs were identi- regions (Table 1). The majority of the genes, classified fied with the Talairach Daemon Client (http://www.talairach by gene ontology analysis as associated with cell adhe- .org). Measures of brain activation (weighted parameter sion, were downregulated in the assayed brain regions estimates) from the suprathreshold cluster were extracted with of persons with SZ.

ARCH GEN PSYCHIATRY/ VOL 69 (NO. 1), JAN 2012 WWW.ARCHGENPSYCHIATRY.COM 10

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 Table 2. Gene Expression Analysis of Selected Nodes of Ranvier Genes in the Superior Temporal Gyrus and Primary Visual Cortexa

STG PVC

Gene Name (Symbol) P Value FC Ratio P Value FC Ratio Neurofascin (NFASC) .02b −1.33 .76 −1.05 Contactin 2 (CNTN2/TAG1) .08 −1.30 .37 −1.27 Ankyrin 3, nodes of Ranvier (ANK3) .01b −1.25 .23 −1.11 Contactin 1 (CNTN1) .30 −1.09 .76 1.03 Contactin-associated protein-like 2 (CASPR2) .19 −1.23 .12 −1.25 Neuronal cell adhesion molecule (NRCAM ) .01b −1.30 .51 1.09 Tenascin-R (TNR) .24 −1.19 .56 1.13 Gliomedin (GLDN ) .23 −1.23 .77 −1.06 Sodium channel, voltage-gated, type II, ␣ subunit, Nav1.2 (SCN2A) .70 −1.03 .79 −1.02 Sodium channel, voltage-gated, type VIII, ␣ subunit, Nav1.6 (SCN8A) .04b −1.34 .16 −1.24

Abbreviations: FC, fold change; PVC, primary visual cortex; STG, superior temporal gyrus. a Tissue samples from the STG and PVC of control subjects (n = 24) and persons with schizophrenia (n = 22) were analyzed by quantitative polymerase chain reaction. The FC ratios represent the ratio of the geometric means for each gene to the 3 housekeeping genes in SZ/control values (positive) and in control/SZ values (negative). b Differentially changed genes (P Ͻ .05).

A All B AnkG 270 kDa Controls SZ 1.4 C/C 2.0 + AnkG 250 kDa P = .01 P = .03 C/T T/T GAPDH 1.2 1.5 1.0

0.8 P = .03 1.0 0.6 Gene Expression 0.4

ANK3 0.5 0.2 Relative Expression of Ank3 Protein

0.0 0.0 n = 24 n = 13 n = 10 n = 22 n = 10 n = 11 Controls SZ Controls SZ

Figure 1. ANK3 gene expression and protein levels are significantly decreased in the superior temporal gyrus of patients with schizophrenia (SZ) compared with control subjects. A, Gene expression of ANK3 is significantly decreased in the superior temporal gyrus of persons with SZ vs control subjects (P=.01; fold change=−1.25). The ANK3 rs9804190–associated differences show reduced expression of the ANK3 transcript in persons with SZ homozygous for the C allele in comparison with the T allele carriers (P=.03). Two subjects were not successfully genotyped. Data are expressed as mean±SEM of individual expression values normalized to the geometric mean of 3 housekeeping genes: PGK1, PPIA, and RPLPO. B, The protein level of the AnkG 270-kDa isoform is significantly decreased in the superior temporal gyrus of patients with SZ compared with control subjects (n=8 per group). The anti-AnkG antibody binds also to a less characterized AnkG 250-kDa isoform, which was also downregulated in patients with SZ (PϽ.02). The individual AnkG expression values shown are normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Horizontal lines indicate the group means.

Eight genes known to be involved in the organiza- for contactin 2 (CNTN2 or TAG1)(P=.08; FC=−1.3). tion and maintenance of NOR were selected from those These changes were gene specific given that the mRNA shown in Table 1 for qPCR analysis in an independent levels of other nodal proteins, including contactin 1 and larger sample of well-matched cases (n=22) and con- (CNTN1), contactin-associated protein-like 2 (CNT- trols (n=24). The expression of these genes was as- NAP2), or Nav1.2 sodium channel, ␣ subunit (SCN2A), sessed in the STG and primary visual cortex (eTable 1 did not show significant differences among persons with and eTable 2), brain regions where we found significant SZ and comparison controls (all PϾ.19). Brain region and minimal alterations of genes expression, respec- specificity was evident from the lack of changes in any tively, in our microarray analysis (Table 1). Other rel- of these genes in the primary visual cortex (all PϾ.12). evant genes (eg, Nav1.6) that were expressed at too low Sample pH, postmortem interval, or age of the donors a level for detection by microarray were also included in did not alter the outcome of the analysis when entered this analysis. As shown in Table 2, the STG of this in- into statistical analyses as covariates. dependent cohort of persons with SZ evidenced signifi- All of the persons with SZ included in these studies cantly reduced mRNA expression of ANK3 (P=.01; fold had received antipsychotic medications for many years. change [FC]=−1.25) (Figure 1), neurofascin (NFASC) To test whether the expression of axoglial junction genes (P=.02; FC=−1.33), neuronal cell adhesion molecule is affected by the classic and most frequently prescribed (NRCAM)(P=.01; FC=−1.3), and Nav1.6 sodium chan- antipsychotic medication in this cohort, the expression nel, ␣ subunit (SCN8A)(P=.04; FC=−1.34) and a trend levels of Ank3, Nrcam, Nfasc, and Scn8a were measured

ARCH GEN PSYCHIATRY/ VOL 69 (NO. 1), JAN 2012 WWW.ARCHGENPSYCHIATRY.COM 11

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 in the cerebral cortex of rats receiving long-term treat- There were no differences in demographic characteris- ment with haloperidol and compared with saline vehicle– tics based on ANK3 rs9804190 genotype (eTable 4). treated controls. Three weeks of relatively high-dose (2 Table 3 shows the association of the ANK3 rs9804190 mg/kg/d) treatment with haloperidol did not influence polymorphism with a broad array of phenotypic cogni- the cerebral cortical expression of the examined nodal tive function measures as revealed by QTPHASE analy- genes (all PϾ.50.). sis. The rs9804190 C allele was associated with statisti- Based on the multiple lines of evidence described ear- cally significant (PϽ.01) lower correct responses on the lier supporting decreased ANK3 mRNA expression lev- 2-back and 3-back conditions of the N-back sequential els, the protein levels of AnkG in STG were measured in letter task, fewer completed categories and an increased a subset of individuals with SZ (n=8) vs comparison sub- number of total and perseverative errors during the Wis- jects (n=8) well matched for age, sex, pH, and postmor- consin Card Sorting Test, and lower self-directedness and tem interval from the larger set of subjects used in the higher self-transcendence scores on the Temperament and qPCR studies described earlier (eTable 1). As shown in Character Inventory. Results were not altered when age, Figure 1, the levels of AnkG protein were significantly education, smoking status (cigarettes per day), and IQ decreased (P=.03) in the STG of persons with SZ rela- were entered as covariates. There were no significant as- tive to controls. sociations with baseline acoustic startle, prepulse inhibition of the startle response, and other cognitive tasks ANK3 RS9804190 C ALLELE INCREASES or personality dimensions, with or without covarying for RISK FOR SZ BY AFFECTING age, IQ, and smoking status (cognitive tests and pre- ANK3 EXPRESSION LEVELS pulse inhibition: all PϾ.05). The final experiment used blood oxygenation level– Given the association of ANK3 with BD10-13 and SZ14 in dependent fMRI to examine the effect of the ANK3 genetic studies and the evidence of ANK3 gene and AnkG rs9804190 polymorphism in an independent cohort of protein expression deficits in the brain of persons with healthy volunteers (n=52). Because of the small num- SZ, we determined whether ANK3 genetic variants are ber of TT homozygotes, all carriers of the T variant were associated with ANK3 gene expression in the STG. Be- grouped together for further analysis. Homozygotes of cause of the low frequency of minor (T) allele homozy- the ANK3 rs9804190 risk C allele compared with the T gotes, carriers of the minor allele (C/T and T/T) were allele carriers showed increased activation (as mea- grouped together. A significant main effect of genotype sured by fMRI) in 2 prefrontal clusters centered in the was observed for rs9804190 in the STG with ANK3 gene left inferior frontal gyrus (BA 11/47; x=−36, y=40, z=−2; expression in individuals with SZ (P=.03). Among per- voxels=218; z score=3.34) and the left middle frontal sons with SZ, those who were homozygous for the C al- gyrus (BA 45/46; x=−46, y=30, z=20; voxels=114; z lele had approximately 27% lower levels of ANK3 ex- score=3.15) during performance of the 3-back version pression compared with individuals with at least 1 minor of the N-back working memory task (Figure 2). There T allele (Figure 1). The UNPHASED analysis revealed that was no effect of genotype on demographic or task per- the rs9804190 C allele was associated significantly formance variables or IQ (all PϾ.53) (eTable 5). (P=.006) with lower ANK3 expression in the STG of individuals with SZ; the estimated additive genetic value between the C and T alleles was −4.17. A significant ef- COMMENT fect of the C/C allele in controls was not found (P=.82). No significant effect was observed for other ANK3- These results provide insights into the mechanistic pro- associated single-nucleotide polymorphisms,10-13 cesses and genetic substrates associated with the fre- rs10994336 or rs10761482, in the STG (all PϾ.18) or quently observed myelin- and oligodendrocyte-related any of the ANK3 genotypes in the primary visual cortex deficits in SZ.1,2 They point to an abnormality in the ex- (all PϾ.13). pression of genes and proteins associated with the integ- Based on the association of the ANK3 rs9804190 C al- rity of the NOR and by extension with the efficiency of lele with lower ANK3 expression levels and the finding neurotransmission. The gene expression results are of decreased ANK3 expression in SZ, the association of strengthened by replication in multiple cohorts and di- the rs9804190 C allele, which increases the risk for SZ, verse methods and by the observation of a novel genetic was examined in a case-control study. Using a larger post- association of at least 1 of the key proteins of the NOR, mortem cohort of white persons with SZ (n=208) rela- AnkG, with SZ and with cognitive functions known to tive to controls (n=64), the study revealed a significant be impaired in SZ. Although the ANK3 single- association of the ANK3 rs9804190 C allele with SZ (per- nucleotide polymorphism is within a noncoding region muted P=.03; odds ratio=1.77; 95% confidence inter- of the ANK3 gene, the gene expression findings indicate val, 1.08-2.88). an association not only with SZ but also with reduced We further examined whether the ANK3 rs9804190 gene expression in the STG. The genetic association find- C allele is associated with poorer performance in SZ- ings are strengthened by their consistency across mul- related behavioral assays. Therefore, the association of tiple study groups and by the fact that the ANK3 genetic the rs9804190 C allele with performance on a test of cog- association with SZ is also known to be associated with nition sensitive to SZ was investigated in 513 demo- increased risk for BD.10-13 There is evidence suggestive graphically homogeneous, healthy, young, white indi- of etiological convergence between SZ and BD12,18,19 con- viduals without psychopathology (LOGOS cohort27). sistent with these results.

ARCH GEN PSYCHIATRY/ VOL 69 (NO. 1), JAN 2012 WWW.ARCHGENPSYCHIATRY.COM 12

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 Table 3. Genetic Association Analysis of ANK3 rs9804190 Polymorphism in the Learning on Genetics of Schizophrenia Spectrum Cohorta

Phenotype UNPHASEDb Mean (SD)

T Allele Cohen d for C/C vs Task Outcome Variable P Value Add Value C/C Carriers T Allele Carriers N-back 2-Back correct .003c 0.33 2.44 (0.8) 2.64 (0.6) −0.27 3-Back correct Ͻ.001c 0.27 1.96 (1.1) 2.29 (0.9) −0.32 WCST Categories .008c 0.25 5.46 (1.0) 5.68 (0.8) −0.24 Total errors .008c −0.04 6.87 (6.2) 5.55 (5.0) 0.23 NTE .009c −0.12 1.77 (2.2) 1.29 (1.4) 0.24 NPE .02 −0.06 4.17 (3.4) 3.58 (3.1) 0.18 TCI Self-directedness .01c 0.05 17.22 (5.2) 18.76 (4.1) −0.32 Self-transcendence .01c −0.07 5.65 (3.3) 4.87 (3.0) 0.24

Abbreviations: NPE, Nelson nonperseverative; NTE, Nelson-type perseverative; TCI, Temperament and Character Inventory; WCST, Wisconsin Card Sorting Test. a The mean (SD), effect size, and adjusted P values are from a permutation test for association of outcome variables with those that reached significance at P Ͻ .05. The add value represents the estimated additive genetic value for the minor allele relative to the more common allele. A negative value signifies a lower score in the risk allele. b The following outcome variables were included in the UNPHASED analysis: for acoustic startle and prepulse inhibition: baseline startle (in 115-dB pulse-alone trials) and prepulse inhibition in 6 prepulse-pulse trial types (75 dB, 30 milliseconds; 75 dB, 60 milliseconds; 75 dB, 120 milliseconds; 85 dB, 30 milliseconds; 85 dB, 60 milliseconds; 85 dB, 120 milliseconds); for neurocognitive assessment: spatial working memory (between errors, 8-box condition; within errors, 8-box condition; strategy), Stockings of Cambridge (problems solved correctly), rapid visual information processing (A, B), N-back (1-, 2-, and 3-back correct), Stroop Interference Task (interference), WCST score (categories achieved, total errors, cards, NTE, NPE), word list (immediate, short delay, and long delay correct responses); and for personality questionnaires: State-Trait Anxiety Inventory, behavioral inhibition and activation system (reward responsive, drive, fun seeking), Eysenck Personality Questionnaire (psychoticism, extraversion, neuroticism), TCI (novelty seeking, harm avoidance, reward dependence, persistence, self-directedness, cooperativeness, self-transcendence), schizotypy questionnaire (magical thinking, paranoid ideation, unusual experiences). c P Յ .01.

The molecular mechanism(s) of the ANK3 genetic sus- ever, despite significant recent advances, the mecha- ceptibility for SZ contributes to reduced gene and pro- nisms of formation and maintenance of the NOR re- tein expression of ANK3/AnkG, with disease-relevant ana- main poorly understood and the full complement of tomic specificity. The expression of ANK3 and other nodal known genes and proteins that contribute to their as- genes is reduced in the STG, a brain region that has been sembly, maintenance, and function is still incom- repeatedly shown to be vulnerable in SZ, but not in the plete.5,6 One speculative mechanism is that NOR abnor- visual cortex where SZ-dependent abnormalities are rarely malities will affect the conduction velocity, leading to observed.8,36 The current observations derived from a spe- attenuated, slow, abnormally variable or absent action cific evidence-based hypothesis testing approach: pre- potentials.4 Indeed, in an ANK3 knockout mouse model, liminary findings from multiregional microarray expres- the generation and persistence of action potentials were sion studies of oligodendrocyte- and myelin-associated dramatically inhibited.37 Such abnormalities will limit the deficits in SZ led to hypothesized abnormalities of the distance over which cortical neurons can fire in syn- NOR. This hypothesis was tested and supported by in- chrony, affecting the normal long-range ␥-oscillations be- dependent gene and protein expression experiments. The tween cortical subfields that could lead to attention defi- observation of abnormal expression of known NOR- cits.4 enriched genes and proteins led directly to genetic asso- AnkG and many of the proteins whose mRNA expres- ciation studies and linkage of genetic association with cog- sions were studied here are not exclusively localized to nitive, behavioral, and neuroimaging intermediate the NOR. AnkG is also essential for the localization of phenotypes repeatedly linked to SZ. many membrane proteins to the axon initial segment,38 The mRNA expressions of the NOR-enriched pro- including the Nav channels that are required for action teins ANK3, NFASC, NRCAM, SCN8A, and CNTN2 were potential generation.37 Cruz et al39 showed that the den- all significantly downregulated in the STG of persons with sity of the AnkG-immunoreactive axon initial segment SZ. Increasing evidence points to a pivotal role of AnkG in subjects with SZ was significantly reduced in super- in the initial assembly and concentration of Nav chan- ficial, but not in deep, cortical layers compared with both nels and cell adhesion molecules to central nervous sys- healthy comparison subjects and subjects with major de- tem NOR.5,6 In addition, a significant role for cell adhe- pressive disorder. This might provide an additional mecha- sion molecules in Nav channels clustering has been nism in which AnkG and Nav clustering abnormalities demonstrated in in vitro experiments where treatments impair action potential generation, disrupting neurotrans- with antibodies against the cell adhesion molecules lead mission and the capacity of some neurons to fire in the to a failure of both Nav channels and AnkG to accumu- repetitive and synchronous fashion required for cortical late at the NOR.5,6 That many of these proteins demon- network oscillations. strated to be crucial to the normal functioning of the NOR While the ANK3 single-nucleotide polymorphisms and to saltatory conduction are abnormally expressed in identified are within noncoding regions of the gene, they SZ and the genetic linkage of ANK3 with SZ attest to their nevertheless support a role for ANK3 in susceptibility to importance to the neurobiology of the disease. How- SZ and performance in cognitive measures. Although our

ARCH GEN PSYCHIATRY/ VOL 69 (NO. 1), JAN 2012 WWW.ARCHGENPSYCHIATRY.COM 13

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 tribute to the detected gene expression abnormalities of A IFG MFG the axoglial genes. We did not find an altered expression level in the cerebral cortex of rats treated with haloperidol. However, the possibility of antipsychotic medication exposure confounds cannot be entirely ruled out, as haloperidol exposure in rats might not exactly mimic the neuroleptic-induced changes in the brain of patients with SZ. Nevertheless, never-medicated, healthy, and asymptomatic ANK3 risk allele carriers present with B Effect of genotype CC> (CT + TT) poorer working memory and executive function, indi- 0.6 ces of reduced functionality (self-directedness), and in-

0.5 creased prefrontal activity during a working memory task. These findings indicate a penetrant mechanism of the 0.4

d ANK3 gene in behavior that might increase the risk for SZ through poor cognition. Collectively, these findings 0.3 identify SZ-associated abnormalities in NOR genes and

Effect Size, 0.2 proteins and provide significant evidence for neurobiological mechanisms and functions that might be key to 0.1 the pathophysiology of SZ and to the development of tar- geted therapeutic interventions. 0.0 IFG MFG

Figure 2. Regional brain activation during 3-back vs 0-back contrast. A, Homozygotes of the ANK3 rs9804190 risk allele (CC; n=31) show Submitted for Publication: March 18, 2011; final revi- significantly increased activation compared with carriers of the T allele sion received May 20, 2011; accepted June 18, 2011. (CT/TT; n=21) in 2 clusters within the inferior frontal gyrus (IFG) and middle frontal gyrus (MFG) (whole-brain analysis, PϽ.05 with false discovery rate Author Affiliations: Department of Psychiatry, The Mount correction). B, Effect size of the mean activation levels of the respective Sinai School of Medicine, New York (Drs Roussos, Kat- suprathreshold clusters. sel, Davis, Siever, and Haroutunian), and James J. Peters Veterans Affairs Medical Center, Bronx (Drs Roussos, strategy of a deductive, direct, postmortem neurobiol- Siever, and Haroutunian), New York; Department of Psy- ogy study minimized serendipitous findings, it did in- chiatry and Behavioral Sciences, Faculty of Medicine, Uni- volve multiple statistical tests. Thus, the potential for spu- versity of Crete, Heraklion, Crete, Greece (Drs Roussos, rious associations because of multiple testing is an Bitsios, and Giakoumaki); and Section of Neurobiology important consideration. However, a consistent pattern of Psychosis (Drs Jogia and Frangou) and Social, Ge- of ANK3 rs9804190 association was found in healthy sub- netic, and Developmental Psychiatry Research Centre jects, involving cognition and functional imaging, and (Ms Rozsnyai and Dr Collier), Institute of Psychiatry, in ANK3 gene and protein expression levels in brain tis- King’s College London, England. sue. The likelihood that the same ANK3 risk allele was Published Online: September 5, 2011. doi:10.1001 consistently associated with illness or with illness- /archgenpsychiatry.2011.110 associated abnormalities across multiple geographically Correspondence: Vahram Haroutunian, PhD, Depart- diverse samples by chance seems small. ment of Psychiatry, The Mount Sinai School of Medi- ANK3 has recently been shown to be a susceptibility cine, Room 4F-33, 130 W Kingsbridge Rd, Bronx, NY gene for BD10,11 and SZ,14 further supporting the notion 10468 ([email protected]). of shared genetic vulnerabilities between BD and SZ.12,18,19 Author Contributions: Drs Roussos and Katsel contrib- However, to our knowledge, there has been no previous uted equally to this work. Drs Roussos and Haroutu- report of an association of the ANK3 rs9804190 variant nian had full access to all of the data in the study and with SZ. This may reflect the fact that P values reported take responsibility for the integrity of the data and the in this a priori hypothesis–driven postmortem brain- accuracy of the data analysis. based genetic association study would not be significant Financial Disclosure: None reported. after genome-wide association study correction for hy- Funding/Support: This work was supported by grants pothesis-free multiple testing. MH066392 and MH064673 from the National Insti- The absence of an ANK3 genetic variant effect on ANK3 tutes of Health (Dr Haroutunian), a Veterans Affairs Men- mRNA expression levels in healthy individuals might be tal Illness Research Education and Clinical Center award due to a lack of power. In patients with SZ, additive ef- (Dr Siever), and a Veterans Affairs merit award (Dr Ha- fects such as epistasis or environmental factors might aug- routunian). ment the effect size of ANK3 genetic variants. We iden- Previous Presentations: This paper was presented at the tified an effect of the rs9804190 C allele in a larger 13th International Congress on Schizophrenia Re- population of healthy subjects (LOGOS cohort) and using search; April 4, 2011; Colorado Springs, Colorado; and a sensitive approach (neuroimaging cohort), further sup- the 66th Annual Meeting of the Society of Biological Psy- porting the notion of inadequate power. chiatry; May 14, 2011; San Francisco, California. Antipsychotic medication exposure is a confounding Online-Only Material: The eTables are available at http: factor in postmortem studies of SZ and therefore can con- //www.archgenpsychiatry.com.

ARCH GEN PSYCHIATRY/ VOL 69 (NO. 1), JAN 2012 WWW.ARCHGENPSYCHIATRY.COM 14

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 wide significant susceptibility loci for schizophrenia and bipolar disorder re- REFERENCES ported to date cross-traditional diagnostic boundaries. Hum Mol Genet. 2011; 20(2):387-391. 1. Davis KL, Stewart DG, Friedman JI, Buchsbaum M, Harvey PD, Hof PR, Bux- 19. Lichtenstein P, Yip BH, Bjo¨rk C, Pawitan Y, Cannon TD, Sullivan PF, Hultman baum JD, Haroutunian V. White matter changes in schizophrenia: evidence for CM. Common genetic determinants of schizophrenia and bipolar disorder in Swed- myelin-related dysfunction. Arch Gen Psychiatry. 2003;60(5):443-456. ish families: a population-based study. Lancet. 2009;373(9659):234-239. 2. Konrad A, Winterer G. Disturbed structural connectivity in schizophrenia pri- 20. Haroutunian V, Perl DP, Purohit DP, Marin D, Khan K, Lantz M, Davis KL, Mohs mary factor in pathology or epiphenomenon? Schizophr Bull. 2008;34(1):72- RC. Regional distribution of neuritic plaques in the nondemented elderly and sub- 92. jects with very mild Alzheimer disease. Arch Neurol. 1998;55(9):1185-1191. 3. Peirce TR, Bray NJ, Williams NM, Norton N, Moskvina V, Preece A, Haroutunian 21. Haroutunian V, Purohit DP, Perl DP, Marin D, Khan K, Lantz M, Davis KL, Mohs V, Buxbaum JD, Owen MJ, O’Donovan MC. Convergent evidence for 2Ј,3Ј- RC. Neurofibrillary tangles in nondemented elderly subjects and mild Alzheimer cyclic nucleotide 3Ј-phosphodiesterase as a possible susceptibility gene for disease. Arch Neurol. 1999;56(6):713-718. schizophrenia. Arch Gen Psychiatry. 2006;63(1):18-24. 22. Haroutunian V, Davies P, Vianna C, Buxbaum JD, Purohit DP. Tau protein ab- 4. Nave KA. Myelination and support of axonal integrity by glia. Nature. 2010;468 normalities associated with the progression of Alzheimer disease type dementia. (7321):244-252. Neurobiol Aging. 2007;28(1):1-7. 5. Poliak S, Peles E. The local differentiation of myelinated axons at nodes of Ranvier. 23. Haroutunian V, Katsel P, Schmeidler J. Transcriptional vulnerability of brain re- Nat Rev Neurosci. 2003;4(12):968-980. gions in Alzheimer’s disease and dementia. Neurobiol Aging. 2009;30(4):561- 6. Susuki K, Rasband MN. Molecular mechanisms of node of Ranvier formation. 573. Curr Opin Cell Biol. 2008;20(6):616-623. 24. Katsel P, Davis KL, Li C, Tan W, Greenstein E, Kleiner Hoffman LB, Haroutunian 7. Katsel P, Davis KL, Haroutunian V. Variations in myelin and oligodendrocyte- V. Abnormal indices of cell cycle activity in schizophrenia and their potential as- related gene expression across multiple brain regions in schizophrenia: a gene sociation with oligodendrocytes. Neuropsychopharmacology. 2008;33(12): ontology study. Schizophr Res. 2005;79(2-3):157-173. 2993-3009. 8. Katsel P, Davis KL, Gorman JM, Haroutunian V. Variations in differential gene 25. Dracheva S, Davis KL, Chin B, Woo DA, Schmeidler J, Haroutunian V. Myelin- expression patterns across multiple brain regions in schizophrenia. Schizophr associated mRNA and protein expression deficits in the anterior cingulate cor- Res. 2005;77(2-3):241-252. tex and hippocampus in elderly schizophrenia patients. Neurobiol Dis. 2006; 9. Katsel P, Dracheva S, Copland CM, Gorman JM, Davis KL, Haroutunian V. Tight 21(3):531-540. and adherens junctions genes of myelinating oligodendrocytes are downregu- 26. Haroutunian V, Katsel P, Dracheva S, Davis KL. The human homolog of the QKI lated in schizophrenia. Biol Psychiatry. 2004;55(8)(suppl):159. gene affected in the severe dysmyelination “quaking” mouse phenotype: down- 10. Ferreira MA, O’Donovan MC, Meng YA, Jones IR, Ruderfer DM, Jones L, Fan J, regulated in multiple brain regions in schizophrenia. Am J Psychiatry. 2006; Kirov G, Perlis RH, Green EK, Smoller JW, Grozeva D, Stone J, Nikolov I, Cham- 163(10):1834-1837. bert K, Hamshere ML, Nimgaonkar VL, Moskvina V, Thase ME, Caesar S, Sachs 27. Roussos P, Giakoumaki SG, Adamaki E, Bitsios P. The influence of schizophrenia- GS, Franklin J, Gordon-Smith K, Ardlie KG, Gabriel SB, Fraser C, Blumenstiel B, related neuregulin-1 polymorphisms on sensorimotor gating in healthy males. Defelice M, Breen G, Gill M, Morris DW, Elkin A, Muir WJ, McGhee KA, William- Biol Psychiatry. 2011;69(5):479-486. son R, MacIntyre DJ, MacLean AW, St CD, Robinson M, Van Beck M, Pereira 28. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured Clinical Interview for AC, Kandaswamy R, McQuillin A, Collier DA, Bass NJ, Young AH, Lawrence J, DSM-IV-TR Axis I Disorders. New York: Biometrics Research Dept, New York Ferrier IN, Anjorin A, Farmer A, Curtis D, Scolnick EM, McGuffin P, Daly MJ, Cor- State Psychiatric Institute; 2002. vin AP, Holmans PA, Blackwood DH, Gurling HM, Owen MJ, Purcell SM, Sklar 29. Maxwell ME. Manual for the FIGS. Bethesda, MD: Clinical Neurogenetics Branch, P, Craddock N; Wellcome Trust Case Control Consortium. Collaborative genome- Intramural Research Program, National Institute of Mental Health; 1992. wide association analysis supports a role for ANK3 and CACNA1C in bipolar 30. Wechsler D. Wechsler Adult Intelligence Scale. 3rd ed. San Antonio, TX: Psy- disorder. Nat Genet. 2008;40(9):1056-1058. chological Corp; 1997. 11. Schulze TG, Detera-Wadleigh SD, Akula N, Gupta A, Kassem L, Steele J, Pearl J, 31. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and Strohmaier J, Breuer R, Schwarz M, Propping P, No¨then MM, Cichon S, Schu- powerful approach to multiple testing. J R Stat Soc Series B Methodol. 1995; macher J, Rietschel M, McMahon FJ; NIMH Genetics Initiative Bipolar Disorder 57(1):289-300. Consortium. Two variants in Ankyrin 3 (ANK3) are independent genetic risk fac- 32. Dudbridge F. Pedigree disequilibrium tests for multilocus haplotypes. Genet tors for bipolar disorder. Mol Psychiatry. 2009;14(5):487-491. Epidemiol. 2003;25(2):115-121. 12. O’Donovan MC, Craddock NJ, Owen MJ. Genetics of psychosis: insights from 33. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, Maller J, views across the genome. Hum Genet. 2009;126(1):3-12. Sklar P, de Bakker PI, Daly MJ, Sham PC. PLINK: a tool set for whole-genome 13. Huang J, Perlis RH, Lee PH, Rush AJ, Fava M, Sachs GS, Lieberman J, Hamilton association and population-based linkage analyses. Am J Hum Genet. 2007; SP, Sullivan P, Sklar P, Purcell S, Smoller JW. Cross-disorder genomewide analy- 81(3):559-575. sis of schizophrenia, bipolar disorder, and depression. Am J Psychiatry. 2010; 34. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Prin- 167(10):1254-1263. cipal components analysis corrects for stratification in genome-wide associa- 14. Athanasiu L, Mattingsdal M, Ka¨hler AK, Brown A, Gustafsson O, Agartz I, Gieg- tion studies. Nat Genet. 2006;38(8):904-909. ling I, Muglia P, Cichon S, Rietschel M, Pietila¨inen OP, Peltonen L, Bramon E, 35. Pavlidis P, Qin J, Arango V, Mann JJ, Sibille E. Using the gene ontology for mi- Collier D, Clair DS, Sigurdsson E, Petursson H, Rujescu D, Melle I, Steen VM, croarray data mining: a comparison of methods and application to age effects in Djurovic S, Andreassen OA. Gene variants associated with schizophrenia in a Nor- human prefrontal cortex. Neurochem Res. 2004;29(6):1213-1222. wegian genome-wide study are replicated in a large European cohort. J Psychi- 36. Haroutunian V, Davis KL. Neuropathology of schizophrenia. In: Breier A, Tran atr Res. 2010;44(12):748-753. PV, Herrera J, Bymaster F, Tollefson GD, eds. Current Issues in the Psychophar- 15. Tkachev D, Mimmack ML, Ryan MM, Wayland M, Freeman T, Jones PB, Starkey macology of Schizophrenia. Baltimore, MD: Lippincott Williams & Wilkins; 2000: M, Webster MJ, Yolken RH, Bahn S. Oligodendrocyte dysfunction in schizophre- 57-70. nia and bipolar disorder. Lancet. 2003;362(9386):798-805. 37. Zhou D, Lambert S, Malen PL, Carpenter S, Boland LM, Bennett V. AnkyrinG is 16. Benes FM, Lim B, Matzilevich D, Walsh JP, Subburaju S, Minns M. Regulation required for clustering of voltage-gated Na channels at axon initial segments and of the GABA cell phenotype in hippocampus of schizophrenics and bipolars. Proc for normal action potential firing. J Cell Biol. 1998;143(5):1295-1304. Natl Acad Sci U S A. 2007;104(24):10164-10169. 38. Rasband MN. The axon initial segment and the maintenance of neuronal polarity. 17. Todtenkopf MS, Vincent SL, Benes FM. A cross-study meta-analysis and three- Nat Rev Neurosci. 2010;11(8):552-562. dimensional comparison of cell counting in the anterior cingulate cortex of schizo- 39. Cruz DA, Weaver CL, Lovallo EM, Melchitzky DS, Lewis DA. Selective alterations phrenic and bipolar brain. Schizophr Res. 2005;73(1):79-89. in postsynaptic markers of chandelier cell inputs to cortical pyramidal neurons 18. Williams HJ, Craddock N, Russo G, Hamshere ML, Moskvina V, Dwyer S, Smith in subjects with schizophrenia. Neuropsychopharmacology. 2009;34(9):2112- RL, Green E, Grozeva D, Holmans P, Owen MJ, O’Donovan MC. Most genome- 2124.

ARCH GEN PSYCHIATRY/ VOL 69 (NO. 1), JAN 2012 WWW.ARCHGENPSYCHIATRY.COM 15