Gene Expression Profiling in Brodmann's Area 46 from Subjects with Schizophrenia

Abstract

Objective: To identify altered gene expression in the dorsolateral prefrontal cortex obtained after death from subjects with schizophrenia.

Method: Restriction fragment differential display (RFDD) was used to measure levels of mRNA in Brodmann area (BA) 46 from schizophrenia and control subjects. Findings on specific mRNA identified with RFDD were further investigated using real-time polymerase chain reaction (real-time PCR), PCR and western blotting.

Results: Levels of mRNA for 63 of approximately 12 500 genes differed in BA 46 in schizophrenia. Subsequent real-time PCR has shown that mRNA for muscleblind protein 1 (MBNL1) and protocadherin 17 (PCDH17) are increased in BA 46 from subjects with schizophrenia of short, but not long, duration. Altered levels of mRNA for neither gene were present in BA 9 from subjects with schizophrenia or in either cortical area from subjects with bipolar 1 disorder. By contrast, both RFDD and real-time PCR failed to show altered expression of the schizophrenia candidate gene disrupted in schizophrenia 1 (DISC1) BA46 from any diagnostic cohort.

Conclusion: The present study has identified genes that are differentially expressed in BA 46 in schizophrenia. Initial studies have shown that there is a need for a careful validation of genes shown to be affected in schizophrenia using high-throughput technologies. In addition the present study has shown that gene expression may vary considerably depending on the duration of schizophrenia. This raises the hypothesis that changing gene expression may be underlying the change in symptom profile that occurs with disease progression in some subjects with schizophrenia.

Keywords

bipolar 1 disorder disrupted in schizophrenia 1 muscleblind protein protocadherin 17 schizophrenia

Schizophrenia is predicted to manifest in individuals with a genetic predisposition who have been exposed to as yet unidentified environmental factors [1, 2]. The outcomes of such genetic and environmental interactions will be complex and would presumably act to alter central nervous system (CNS) function in a way that causes the distinctive symptoms of schizophrenia [3]. It has long been proposed that changes in gene expression in the CNS are likely to contribute to abnormal brain function in subjects with schizophrenia [4]. Thus, to accelerate the identification of changes in gene expression, a number of studies have measured levels of mRNA in post-mortem CNS from subjects with the disorder using complimentary DNA (cDNA) microarrays. The power of this approach is that levels of thousands of mRNAs can be measured simultaneously.

The first outcome from microarrays studies has been a reinforcement of the hypothesis that schizophrenia is a complex disorder. These studies also suggest that schizophrenia involves changes in presynaptic functioning [5, 6], regulators of G-protein signalling [7, 6], metabolic pathways [8–10], γ-aminobutyric acid neurons [11], ionotropic glutamate receptors [6, 12, 13], myelination [14] and protein turnover [10]. In addition, some studies have focused attention on abnormalities in expression of individual genes such as tyrosine 3-monoxygenase/tryptophan 5 monoxygenase [15] and apolipoprotein L1 [16]. Another finding arising from microarray studies is that changes in levels of individual mRNAs can display both regional [15] and diagnostic [16] specificity. Together, and as more comprehensively reviewed elsewhere [17], these data suggest that there is an ongoing need to continue to use high-throughput methodologies to identify changes in gene expression in schizophrenia.

High-throughput screening techniques for mRNA in tissue can be divided into open and closed system approaches [18]. The closed system approaches, such as the microarrays, utilize information on specific gene sequences to allow the generation of probes that can hybridize to specific mRNA that are transcribed from genes of interest. While the closed system approaches have advantages with regards to reproducibility and speed of gene identification, they are limited by sensitivity (one part in 10⁵ or 10⁴ depending on the nucleotide sequence) and cost (multiple high-intensity arrays per study). The alternative open system approaches, which are polymerase chain reaction (PCR)-based methodologies, detect mRNA based on nucleotide sequence using adaptors that can be attached to multiple DNA fragments, allowing these fragments to be amplified at one time. The open system approach is problematic in controlling for inter-assay variability but has the advantage of increased sensitivity (one part in 10⁶), allowing the study of mRNA with lower copy number.

Given that genes expressed at low abundance have been shown to be important in the pathology of schizophrenia [6], we decided to use an open system approach, restriction fragment differential display (RFDD), to measure levels of mRNA in the dorsolateral prefrontal cortex (DLPFC) from subjects with schizophrenia and age/sex-matched controls. Studies were undertaken in the DLPFC because there is a large body of data showing that the functioning of this CNS region is affected in subjects with schizophrenia [19]. In addition, we have used real-time PCR, PCR and western blot analyses with CNS tissue from relatively large cohorts of subjects, peripheral tissue and CNS from rats treated with antipsychotic drugs to extend our initial findings from RFDD.

Methods

Tissue collection

Human tissue, clinical follow up and diagnosis

Prior to commencement, approval for this study was obtained from the Ethics Committee of the Victorian Institute of Forensic Medicine and the North Western Mental Health Program Behavioural and Psychiatric Research and Ethics Committee. All tissue utilized in the studies were obtained from the Victorian Brain Tissue Repository in the Rebecca L. Cooper Research Laboratories at the Mental Health Research Institute, Melbourne.

Psychiatric diagnoses were made according to DSM-IV criteria [3] by consensus between two senior psychiatrists and a psychologist after an extensive case history review using a structured diagnostic instrument, the Diagnostic Instrument for Brain Studies (DIBS) [20]. Duration of illness (DOI) was calculated as the time from first contact with a psychiatric service to death, and the final dose of antipsychotic drug (FRADD) was converted to chlorpromazine equivalents. In cases where death was witnessed, the time between death and autopsy was taken as the post-mortem interval (PMI). Where death was not witnessed, tissue was taken only from individuals who had been seen alive up to 5 h before being found dead, in these cases PMI was taken as the interval half way between the donor being found dead and last being seen alive. In all cases, cadavers were refrigerated within 5 h of being found and tissue was rapidly frozen to −70^oC within 30 min of autopsy. The pH of the CNS tissue was measured as described previously [21] because this measure provides a better indication of mRNA integrity than PMI [22].

For RFDD experiments, tissue was obtained from Brodmann area (BA) 46 from the left hemisphere of 10 subjects with a DSM IV diagnosis of schizophrenia and from the same region from 10 subjects with no history of psychiatric illness (controls; Table 1). To avoid bias, the subjects with schizophrenia were not selected using any predetermined criteria. The control subjects were matched for sex and closely matched for age to the subjects with schizophrenia and thus there was no significant difference in age (p = 0.82), PMI (p = 0.59) or brain pH (p = 0.14) between the diagnostic cohorts (Table 1).

Table 1.

Pooled mRNA cohorts for RFDD

	Schizophrenia	Control
Age (years)	45.8±5.3	47.6±5.6
Sex ratio (M/F)	5/5	5/5
PMI (h)	38.1±6.2	43.2±6.9
FRADD†	503±100	NA
pH	6.43±0.04	6.35±0.04
DOI (years)	11.5±2.4	NA

DOI, duration of illness; FRADD, final dose of antipsychotic drug; NA, not applicable; PMI, post-mortem interval; RFDD, restriction fragment differential display

†Chlorpromazine equivalents per day.

Subsequently, for the experiments using real-time PCR, BA 46 and BA 9 were obtained from the left hemisphere of 30 subjects with schizophrenia, eight subjects with bipolar 1 disorder (BPD1) and 30 control subjects matched for age, sex, and PMI (Table 2). Thus, once again there were no significant differences in age (F = 2.33, df = 2,67, p = 0.1), brain pH (F = 0.26, df = 2,67, p = 0.77) or PMI (F = 0.10, df = 2,67, p = 0.90) between the three cohorts of subjects. In addition, for the disease cohorts, there were no significant differences in DOI (p = 0.99) or FRADD (p = 0.84) between subjects with schizophrenia and BPD1 (Table 2). RNA from the subjects used in the RFDD experiments were included in the cohorts used for real-time PCR.

Table 2.

Differential display validation study cohorts

	Schizophrenia	Bipolar 1 disorder	Controls
n	30	8	30
Age (years)	43.8±3.4	58.9±3.6	44.8±3.4
Sex ratio (M/F)	24/6	4/4	24/6
DOI (years)	18.9±3.0	18.8±5.0	NA
FRADD†	559±110	620±345	NA
PMI (h)	42.1±2.5	40.1±5.0	42.7±2.7
pH	6.32±0.03	6.27±0.10	6.33±0.04

DOI, duration of illness; FRADD, final dose of antipsychotic drug; NA, not applicable; PMI, post-mortem interval

†Chlorpromazine equivalents per day.

The cohorts of subjects with schizophrenia used for real-time PCR were made up of individuals with a short (<7 years) and long (>22 years) DOI (Table 3). Dividing the schizophrenia subjects and their matched controls according to DOI of the schizophrenia subjects (short-duration schizophrenia + their age- and sex-matched controls, SDOI; long-duration schizophrenia + their age- and sex-matched controls, LDOI) resulted in there being a significant variance in age between the subjects with SDOI and LDOI (F = 97.1, df = 1, 56, p < 0.0001) but not between the diagnostic cohorts (F = 0.1, df = 1,56, p = 0.85) and there was no interaction between the variables (F = 0.03, df = 1,56, p = 0.86; Table 3). Predictably, the variation with age was due to the SDOI subjects (29.6±2.2 years) being significantly younger than the subjects with LDOI (59.0±1.9 years, p < 0.0001). In the case of PMI, there was significant variance associated with SDOI and LDOI (F = 5.3, df = 1,56, p < 0.05) but not with diagnosis (F = 0.02, df = 1,56, p = 0.87) and there was no interaction between the variables (F = 0.06, df = 1,56, p = 0.81). The variance in PMI was due to the subjects with SDOI (46.6±2.2 h) having longer PMI than the LDOI subjects (38.3±2.7 h, p = 0.02). There was no significant variance in brain pH with either DOI (F = 1.5, df = 1,56, p = 0.23) or diagnosis (F = 0.05, df = 1,56, p = 0.82). For the subjects with schizophrenia, there was no significant difference in FRADD (p = 0.26) between SDOI and LDOI subjects.

Table 3.

Duration of illness study cohorts

	DOI < 7 years		DOI > 22 years
	Schizophrenia	Control	Schizophrenia	Control
Age (years)	29.4±3.2	29.9±3.1	58.2±2.8	59.7±2.7
Sex ratio (M/F)	12/3	12/3	12/3	12/3
DOI (years)	4.1±0.5	NA	33.7±2.5	NA
FRADD†	684±200	NA	434±89	NA
PMI (h)	45.8±3.4	47.3±3.1	38.4±3.6	38.1±4.3
pH	6.29±0.04	6.30±0.05	6.35±0.04	6.36±0.04

DOI, duration of illness; FRADD, final dose of antipsychotic drug; NA, not applicable; PMI, post-mortem interval

†Chlorpromazine equivalents per day.

Post-mortem CNS from various regions from three other subjects with no history of psychiatric or neurological history was also collected for the study of the muscleblind protein 1 (MBNL1) expression in the human brain.

Rat tissue

Cortical tissue was obtained from male rats that had been treated for 3 months with vehicle, or the antipsychotic drug chlorpromazine, at a range of doses (0.1, 1.0 and 10 mg kg⁻¹ day⁻¹: five rats per treatment group), in drinking water for 3 months. This treatment duration was to ensure that rats had received antipsychotic drugs for a time greater than the minimum duration required for the drugs to have clinically significant effects in humans [23]. For tissue expression studies, tissues were collected from three male and three female rats between 6 and 10 weeks weighing between 200 and 250 g. In both cases, the tissue was rapidly frozen in isopentane in dry ice prior to storage at −70^oC until required.

RNA purification and cDNA template preparation

For RFDD, real-time PCR and PCR, total RNA was extracted using Trizol reagent (Invitrogen Australia Pty Ltd, Mount Waverley, VIC, Australia).

For RFDD, RNA samples from 10 schizophrenia and 10 control subjects were pooled to give two samples, one for each diagnostic cohort. First-strand cDNA from each of the two pools was synthesized using 1 g total RNA, an oligo-dT₂₅ primer (12.5 µmol L⁻¹) and displayTHERMO reverse transcriptase (100 U; Display Systems Biotech, Inc., Vista, CA, USA) in 25 µL volume by incubating at 42°C for 2 h. Second-strand cDNA was synthesized at 16°C for 2 h. The resulting cDNA was digested with a restriction endonuclease (TaqI) at 65°C for 2 h. Specific adaptors were then ligated to the ends of the cDNA fragments at 37°C for 3 h to give cDNA templates for use in the RFDD assay.

Restriction fragment differential display

The RFDD analysis was carried out as described in detail in the displayPROFILE kit (Display Systems Biotech). Briefly, the cDNA templates were amplified in 64 PCR reactions consisting of a fluorescently labelled 5′ primer and one of 64 specific display PROBE 3′ primers (Display Systems Biotech) that anneal to the junction between the linker and cDNA fragment. Three bases (N₁N₂N₃, where N = A, C, T or G), distinct for each 3′ primer, extend into the cDNA sequence allowing for specific amplification of cDNAs. Each PCR reaction amplifies a different subset of cDNA products and to amplify all of the 4³ variations, 64 different primers were required with each set of primers being run in a separate PCR reaction. All cDNAs generated in a given PCR reaction are referred to as making up an ‘expression window’ and PCR products were generated within 64 expression windows. Notably, the primer design ensured that each cDNA fragment was amplified in two separate expression windows, and the experimenter was blind to the expression windows in which these duplicate amplifications were performed.

For each PCR reaction, an aliquot (0.2 µL) of cDNA template was mixed with 2.0 µL of displayTAQ FL 10× reaction buffer (Display Systems Biotech), 0.8 µL dNTP mix (5 µmol L⁻¹ each), 1.0 µL fluorescently labelled 5′ primer (4 µmol L⁻¹), 0.3 µL displayTAQ FL polymerase and 4.0 µL of displayPROBE3′ primer (1 µmol L⁻¹). Sterile H₂O was added to give a final volume of 20 µL. Each reaction was amplified with touchdown PCR using the following profile: initial denaturing 94^oC for 5 min; and 10 cycles: denaturing, 94°C for 1 min; annealing, 60^oC for 30 s; extension, 72^oC for 30 s during which the annealing temperature is reduced by 0.5°C after each cycle so that the final annealing temperature is 55°C. An additional 35 cycles were performed at a constant annealing temperature (55°C) with a final extension of 72°C for 5 min. Electrophoresis was used to resolve each amplified cDNA so that the length and relative abundance of each DNA fragment could be measured.

Fold changes in levels of each DNA fragment in schizophrenia were calculated as follows:

Increases in expression: (schizophrenia=control) × 100=% increase

Decreases in expression: (control=schizophrenia) ×100=% decrease

This approach meant that the magnitude of increases and decreases in gene expression could be expressed as percent changes without constraining one variable between 0 and 100%. Experiments (data not shown) examining the reproducibility of the system suggested that levels of amplified DNA fragments had to differ by >300% between diagnostic cohorts in one expression window to give a high probability that differential expression had been detected. However, analysis of the completed RFDD data showed that a change in mRNA levels in schizophrenia of 400 or 500% in one experimental window gave a 57 and 64% prediction, respectively, of a change in the second window being >300%. By contrast, a change >600% in one window was >99% predictive of an expression difference of >300% in the second window. Thus, we used two criteria to identify differentially expressed genes. These were that a gene had to be detected in both predicted expression windows and there had to be a >600% difference in mRNA levels between diagnostic cohorts in one expression window. The identity of the genes meeting these criteria involved entering the fragment size and expression window number (the specific trinucleotide sequence used for amplification) in Display Systems Biotech's displayFIT program. This program accesses a proprietary database via the internet to determine gene identity.

Real-time polymerase chain reaction

Real-time PCR was performed using the LightCycler thermal instrument (Roche Diagnostics, Mannheim, Germany). Reactions were performed in 20 µL volume with 0.5 µmol L⁻¹ primers and MgCl₂ concentration optimized between 2 and 5 mmol L⁻¹. Nucleotides, Taq DNA polymerase and buffer were included in the LightCycler DNA Master SYBR Green I mix (Roche Diagnostics). The cDNA from each subject was diluted to 1 in 20 prior to PCR amplification. A typical run consisted of a 10 min denaturation step at 95^oC followed by 40 cycles of 95^oC for 10 s, 55–64^o (primer dependent) annealing for 5 s, and 72^oC extension for 8 s. Detection of the fluorescence was carried out either at the end of the 72^oC extension period or after an additional 3 s step at 2^oC below the product melting temperature. To confirm amplification specificity, the PCR products from each primer pair were subjected to a melting curve analysis as well as subsequent agarose gel electrophoresis and sequence verification. The quantification data were analysed with the LightCycler analysis software.

To normalize expression from real-time PCR, the levels of expression of β-actin mRNA was measured in BA 9 and BA 46 in all samples. Levels of β-actin mRNA did not differ between diagnostic cohorts (F = 1.30, df = 2,130, p = 0.28) or between regions (F = 1.78, df = 1,130, p = 0.18) and therefore could be used as the housekeeping gene for the present study. Thus, gene amplification levels were normalized against β-actin measured against a β-actin standard curve using serial 10-fold dilutions ranging from 10 ng to 0.001 pg by amplification of a 203 bp β-actin fragment, against a template that was prepared by excising from a gel, purifying and sequencing. In normalizing to β-actin the capacity to compare absolute levels of specific mRNA between regions was lost.

MBNL1 expression localization studies

Primer sets to amplify human MBNL1 (forward sequence 5′-GAGATGAATGTGCCAAGCAA–3′; reverse sequence 5′–AGAATGGTGCCCTTTGACTG-3′) and rat (forward sequence 5′-GGCCTTGTTATAAATAAGCTGCAT-3′; reverse sequence 5′-GACTATATACAGAAGTGCAAAAAGTCA-3′) were synthesized by Invitrogen.

PCR reaction was performed in a final volume of 50 µL, using Red Hot Taq DNA Polymerase, 1.25 U, in the buffer supplied by the manufacturer (Invitrogen Australia), 1.5 mmol L⁻¹ MgCl₂, primer mix (forward and reverse) 0.5 µmol L⁻¹ each, 0.8 mmol L⁻¹ dNTP mix, 2 µL from a 1 in 10 dilution of the first-strand cDNA template synthesized as aforedescribed, and one cycle of initial denaturation at 94^oC for 2 min followed by 30 cycles consisting of denaturation (94^oC for 30 s), annealing (60^oC for 30 s) and elongation (72^oC for 30 s), followed by a final elongation period of 5 min at 72^oC. PCR products were separated on an agarose gel and imaged using a BioRad gel doc system (BioRad, Regents Park, NSW, Australia).

MBNL protein analysed using western blotting

Protein homogenates (20 µg) from BA 46 were loaded onto a 4% dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) stacking gel and separated in a 10% SDS-PAGE resolving gel along with BioRad broad range biotinylated standards (molecular weight MW markers: two lanes), a sample of bovine serum albumin (one lane: independent check for molecular weight) and internal control (two lanes: see following). The proteins were separated by maintaining a constant voltage of 150 V until dye moved to the bottom of the gel.

The proteins on each gel were transferred overnight onto Hybond-ECL (GE Health Life Sciences, North Shore, NSW, Australia) at constant 40 mA. The proteins on the nitrocellulose membrane were visualized using a Ponceau S stain to ensure that the transfer had been effective; the nitrocellulose membranes were de-stained by washing in distilled water and the lanes containing the MW markers removed prior to the nitrocellulose membrane being blocked in 0.1% tween-20 in Tris-buffered saline (TTBS) for 1 h at room temperature. After blocking with TTBS, the nitrocellulose membrane were incubated with TTBS containing 1:2000 dilution of a monoclonal anti-MBNL antibody (kind gift from Professor Maurice S. Swanson, Department of Molecular Genetics and Microbiology, University of Florida) in TTBS for 2 h at room temperature. Each nitrocellulose membrane was then washed 3×5 min in TTBS and incubated with a Dako goat anti-mouse Ig antibody (Dako, Botany, NSW, Australia) diluted 1:2000 in TTBS for 2 h at room temperature. At the completion of this incubation each nitrocellulose membrane and the corresponding MW standard strips (that had been removed earlier) were incubated for 2 h at room temperature with avidin–horseradish peroxidase conjugate (diluted 1:3000 in TTBS). All nitrocellulose membranes and MW strips were then washed for 3×5 min in TTBS and then exposed to Amersham ECL (GE Health Life Sciences, North Shore, NSW, Australia) for 1 min at room temperature. The MW strips and nitrocellulose membranes were realigned on the image plate of a Kodak 440CF image station and images captured for 30 s. Image analysis first involved ensuring that the MW standards had been correctly aligned by confirming the MW of the bovine serum albumin on each gel. The density of the MBNL protein in each sample was then measured and expressed as a ratio of the density of the MBNL protein in the internal control (ratio IC) to control for gel-to-gel variation.

Internal control sample

Prior to measuring MBNL in BA 46, sufficient protein homogenate (internal control: IC) was prepared from the cerebellum of a control subject to allow samples to be run on every gel containing BA 46 from the schizophrenia and control subjects. Initially, this homogenate was run in each of 12 wells on four gels (48 samples over all) and the optical density (OD) of each immunopositive MBNL band measured as aforedescribed. Using these data the intra-gel variation was shown to be 9% and the between-gel variation to be 12%. A sample of IC was included in two lanes of each subsequent gel in which protein from BA 46 were separated and the OD of each sample was expressed as a ratio of the IC to control for gel-to-gel variation in our analyses.

Statistics

One-way ANOVA within CNS-region diagnoses was used to identify disease-specific variation with a post-hoc Dunnett's test to identify differences between schizophrenia and control subjects. Two-way ANOVA with post-hoc Bonferroni post-tests were used to identify changes in levels of MBNL protein, age, PMI and brain pH between diagnosis and DOI. t-Tests were utilized to identify possible significant differences in the DOI and FRADD between the cohorts of subjects with schizophrenia and BPD1. Relationships between experimental parameters and potential confounding factors were analysed using Pearson single product moment correlations calculated using an assumed straight line fit. All analyses were completed using GraphPad Prism version 4.0 (GraphPad Software Inc., San Diego, CA, USA).

Results

Studies in human central nervous system

Restriction fragment differential display

RFDD allowed the visualization of approximately 12 500 gene fragments in the pools of mRNA prepared from BA 46 from 10 schizophrenia and 10 control subjects. Using our stated criteria, 62 mRNAs were shown to be present at altered levels in tissue from subjects with schizophrenia (Table 4), with mRNA from 26 genes (41.9%) being present at higher levels. On dividing the differentially expressed genes in BA 46 into families, based on broad functional relationships, genes encoding receptors and membrane transport proteins as well as those involved in signal transduction, cell adhesion, development, transcription, DNA/RNA binding, energy/metabolism and cell structure/matrix seemed particularly affected in schizophrenia. The 62 genes were distributed across most chromosomes (1–19, 21 and X) and the function of one gene is yet to be defined (Table 4).

Table 4.

Genes with altered levels of expression in BA 46 from subjects with schizophrenia

Functional grouping	Official gene name	Official symbol	Gene ID	Gene loci	Fold change
Cell adhesion	Neural cell adhesion molecule 2	NCAM2	4685	21q21.1	7.25
	Membrane-associated guanylate kinase, WW and PDZ domain containing 1	MAGI1	9223	3p14.1	−8.46
	Protocadherin 17	PCDH17	27253	13q14.13	−11.49
	Protocadherin gamma subfamily C, 3	PCDHGC3	5098	5q31	11.39
	Cadherin 20, type 2	CDH20	28316	18q22–q23	14.80
	C-type lectin domain family 2, member B	CLEC2B	9976	12p13–p12	6.28

Signal transduction	Guanine nucleotide binding protein (G protein), alpha activating activity polypeptide O	GNA01	2775	16q13	−5.41
	Phospholipase C, gamma 2 (phosphatidylinositol-specific)	PLCG2	5336	16q.24.1	−5.50
	Regulator of G-protein signalling 6	RGS6	9628	14q24.3	7.89
	Mitogen-activated protein kinase 3	MAPK3	5595	16p12 − p11.2	−11.67
	Protein tyrosine phosphatase, non-receptor type 21	PTPN21	11099	14q31.3	−6.34
	Phospholipase C, gamma 2 (phosphatidylinositol-specific)	PLCG2	5336	16q24.1	6.58
	Naked cuticle homolog 1 (Drosophila)	NKD1	85407	16q12	13.25
	Tyrosine kinase with immunoglobulin-like and EGF-like domains 1	TIE	7075	1p34–p33	−6.88
	A kinase (PRKA) anchor protein 6	AKAP6	9472	14q13.1	7.53
	Neurobeachin	NBEA	26960	13q13	−5.07
	Chromosome 13 open reading frame 18	C13orf18	80183	13q14.11	−5.33
	Phosphatidylinositol 3,4,5-trisphosphate-dependent RAC exchanger 1	PREX1	57580	20q13.3	8.62

Development	Notch homolog 1, translocation-associated (Drosophila)	NOTCH1	4851	9q34.3	10.88
	Neuron navigator 2	NAV2	89797	11q15.1	−16.34
	Nidogen 1	NID	4811	1q43	8.00
	Muscleblind-like (Drosophila)	MBNL1	4154	3q25	33.15
	NCK-associated protein 1	NCKAP1	10787	2q32	−6.13
	Mesoderm specific transcript homolog (mouse)	MEST	4232	7q32	8.60
	Platelet-derived growth factor D	SCDGF-B	80310	11q22.3	−7.22

Transcription	Estrogen-related receptor gamma	ESRRG	2104	1q41	−6.45
	v-myc myelocytomatosis viral oncogene homolog 2 (avian)	MYCL2	4611	Xq22–q23	−6.83
	T-box 18	TBX18	9096	6q14–q15	4.89
	AF-1 specific protein phosphatase	Not Avail	196515	12q24.13	−6.01
	General transcription factor IIIC, polypeptide 1, alpha 220 kDa	GTF3C1	2975	16p12	−12.59
	Polymerase (DNA directed), gamma	POLG	5428	15q25	−5.44

Energy/Metabolism	Phosphoenolpyruvate carboxykinase 2	PCK2	5106	14q11.2	8.31
	Malate dehydrogenase 2, NAD (mitochondrial)	MDH2	4191	7q12.3–q11.2	−5.27
	Hexose-6-phosphate dehydrogenase (glucose 1-dehydrogenase)	H6PD	9563	1p36.13	−4.96
	Cytochrome P450, family 2, subfamily C, polypeptide 18	CYP2C18	1562	10q24	−4.83

Structural/matrix	Fibrillin 2 (congenital contractural arachnodactyly)	FBN2	2201	5q23–q31	−5.58
protein	Collagen, type IX, alpha 2	COL9A2	1298	1p33–p32	−12.39
	Fibulin 2	FBLN2	2199	3p25p24	5.10
	Filamin C, gamma (actin binding protein 280)	FLNC	2318	7q32q35	−7.24

DNA/RNA binding	Cleavage and polyadenylation specific factor 1, 160 kDa	CPFS1	29894	8q24.3	−8.38
Proteins	Splicing factor 3b, subunit 1, 155 kDa	SF3B1	23451	2q33.1	10.40
	Ribosomal protein L12	RPL12	6136	9q34	−6.43
	DDI1, DNA-damage inducible 1, homolog 2 (S. cerevisiae)	DDI2	84301	1p36.13	−5.33
	Structural maintenance of chromosomes 5	SMC5	23137	9q21.11	−13.51

Membrane	Solute carrier family 6 (proline IMINO transporter), member 20	SLC6A20	54716	3p21.3	5.40
Transport	Solute carrier family 12 (sodium/potassium/chloride transporters), member 2	SLC12A2	6558	5q23.3	7.40

Receptors	Glycine receptor, alpha 3	GLRA3	8001	4q33–q34	−6.34
	Gamma-aminobutyric acid (GABA) A receptor, beta 2	GABRB2	2561	5q34	4.21
	GP5 glycoprotein V (platelet)	GP5	2814	3q29	10.94
	Activin A receptor, type IB	ACVR1B	91	12q13	−11.68
	T cell receptor beta variable 2	TRBV2	28620	7q34	12.32

Miscellaneous	Calcium homeostasis endoplasmic reticulum protein	CHERP	10523	19p13.1	−5.93
	Myelin oligodendrocyte glycoprotein	HGNC	4340	6p22–p21.3	−7.86
	Proprotein convertase subtilisin/kexin type 6	PCSK6	5046	15q26.3	10.54
	Exocyst complex component 5	EXOC5	10640	14q22.3	−4.44
	Proline/serine-rich coiled-coil 1	PSRC1	84722	1p13.3	−6.52
	Carbohydrate (chondroitin 4) sulfotransferase 11	CHST11	50515	12q	−8.26
	La ribonucleoprotein domain family, member 5	LARP5	23185	10p15.3	−5.66
	Chromosome 14 open reading frame 34 (C14orf34)	C14orf24	55673	17q23.2	5.37
	Centrosomal protein 135 kDa	CEP135	9662	4q12	−5.66
	Seven in absentia homolog 2 (Drosophila)	SIAH2	6478	3q25	21.81

Unknown	KIAA0913	KIAA0913	23053	10q22.3	9.32

On completion of RFDD three genes were selected for further investigation. Two of these genes, MBNL1 and protocadherin 17 (PCDH17), were chosen because the mRNA from these genes was shown to be robustly increased and deceased in schizophrenia, respectively. The third gene, dissociated in schizophrenia 1 (DISC1), was chosen because it appears to be a strong candidate gene for schizophrenia and because mRNA levels for that gene were not changed according to RFDD [24].

Real-time polymerase chain reaction

There was significant variation in the level of MBNL1 mRNA in BA 46 (F = 9.82, df = 2,67, p = 0.0002) but not in BA 9 (F = 0.74, df = 2,67, p = 0.48) with diagnosis (Figure 1a). Post-hoc analysis showed that the variance in BA 46 was due to a significant increase in the level of MBNL1 mRNA in schizophrenia (mean±SEM: 0.78±0.05 MBNL1/β-actin ratio) compared to that in control subjects (mean±SEM: 0.53±0.04 MBNL1/β-actin ratio, p < 0.01). There was no difference in levels of MBNL1 mRNA in BA 46 from subjects with BPD1 (mean±SEM: 0.41±0.07 MBNL1/β-actin ratio). Further analyses comparing results from SDOI and LDOI subjects showed that there was a significant increase in levels of MBNL1 mRNA in BA 46 from SDOI subjects compared to their matched controls (mean±SEM: 0.87±0.08 vs 0.48±0.06 MBNL1/β-actin ratio, p < 0.0001; Figure 1b). By contrast, MBNL1 mRNA did not differ in BA 46 from LDOI subjects and their age- and sex-matched controls (mean±SEM: 0.69±0.06 vs 0.58±0.07 MBNL1/β-actin ratio, p = 0.26). There were no relationships between levels of MBNL1 mRNA and any subject demographic or tissue collection variables.

Figure 1.

Expression levels (mean±SEM), determined using real-time polymerase chain reaction and normalized to β-actin, of muscleblind protein 1 (MBNL1) in (a) Brodmann area (BA) 46 and BA 9 from 30 subjects with (▪) schizophrenia, eight subjects with bipolar 1 disorder and 30 (□) control subjects and in (b) BA 46 from 15 subjects with schizophrenia short duration of illness (SDOI) <7 years and 15 subjects with schizophrenia long duration of illness (LDOI) >22 years with their appropriately matched controls. ^∗∗p < 0.01; ^∗∗∗p < 0.001.

Levels of PCDH17 mRNA showed significant variation with diagnosis in BA 46 (F = 5.70, df = 2,67, p = 0.005) and BA 9 (F = 3.74, df = 2,67, p = 0.03; Figure 2a). Post-hoc analysis showed that the variance in BA 46 was due to a significant increase in the level of PCDH17 mRNA in the tissue from subjects with schizophrenia (mean±SEM: 2.25±0.32 PCDH17/β-actin ratio) compared to that in tissue from the control subjects (mean±SEM: 1.77±0.19 PCDH17/β-actin ratio: p < 0.05). Despite a lower mean, levels of PCDH17 mRNA did not differ significantly in BA 46 from subjects with BPD1 (mean±SEM: 0.79±0.18 PCDH17/β-actin ratio; p > 0.05) compared to controls. Although global analysis showed that levels of PCDH17 mRNA varied with diagnosis in BA 9, post-hoc analysis did not identify any significant variation between levels of mRNA in tissue from the three cohorts of subjects (mean±SEM: controls = 0.20±0.02 PCDH17/β-actin ratio; schizophrenia = 0.33±0.04 PCDH17/β-actin ratio and BPD1 = 0.15±0.04 PCDH17/β-actin ratio). Hence the global variance in mRNA reflected trends to increased PCDH17 mRNA in schizophrenia and decreased PCDH17 mRNA in BPD1. Significantly, the results from real-time PCR were in the opposite direction to what was predicted from RFDD, in which levels of PCDH17 mRNA were shown to be decreased.

Figure 2.

Expression levels (mean±SEM), determined using real-time polymerase chain reaction analysis and normalized to β-actin, of protocadherin 17 (PCDH17) in (a) Brodmann area (BA) 46 and BA 9 from 30 subjects with (▪) schizophrenia, eight subjects with bipolar 1 disorder and 30 (□) control subjects and in (b) BA 46 from 15 subjects) with schizophrenia short duration of illness (SDOI) <7 years and 15 subjects with schizophrenia long duration of illness (LDOI) with their appropriately matched controls. ^∗p < 0.05; ^∗∗p < 0.01.

Further analyses, examining the potential temporal effects of DOI, showed that there was a significant increase in levels of PCDH17 mRNA in BA 46 from SDOI subjects compared to their matched controls (mean±SEM: 0.42±0.06 vs 0.21±0.03 PCDH17/β-actin ratio, p < 0.01; Figure 2b). In contrast, levels of mRNA in BA 46 from LDOI subjects (mean±SEM: 0.24±0.03 PCDH17/β-actin ratio) did not vary significantly (p = 0.38) from that in control tissue (mean±SEM: 0.20±0.02 PCDH17/β-actin ratio). There were no relationships between levels of PCDH17 mRNA and any demographic or tissue collection variables.

DISC1 mRNA was measured only in BA 46, where there was no significant variation in the levels of mRNA with diagnosis (mean±SEM: controls = 0.67±0.14 vs schizophrenia = 0.62±0.2 vs BPD1 = 1.1±0.37 DISC1/β-actin ratio, F = 0.66, df = 2,67, p = 0.52). A comparison of DISC1 mRNA in DOI did not show any variance with diagnosis in the SDOI (F = 0.04, df = 1,56, p = 0.83) or LDOI (F = 0.14, df = 1,56, p = 0.7) cohorts. There were no relationships between levels of DISC1 mRNA and possible confounding variables.

MBNL1 mRNA distribution in humans and rats

Studies using PCR and human post-mortem CNS showed abundant levels of MBNL1 mRNA in BA 46 and BA 40 in tissue from three control subjects (Figure 3). At a qualitative level, MBNL1 mRNA was less abundant in BA 9, hippocampus, caudate, putamen, thalamus and amygdala from the same donors. In addition, a qualitative study using RNA prepared from three male and three female rats showed MBNL1 mRNA in CNS, kidney, muscle, stomach, lung, heart, ovary, testis, liver and spleen.

Figure 3.

Distribution of muscleblind protein 1 (MBNL1) in tissue from various regions of the (a) human central nervous system (CNS) and (b) different rat tissues; the data shown are from one human subject and one rat but are typical of data from all human subjects and rats studied. Polymerase chain reaction (PCR) was performed on first-strand cDNA prepared from human caudate, putamen, thalamus, amygdala, hippocampus, Brodmann area (BA) 40, 9 and 46 as well as in rat kidney, muscle, stomach, ovaries/testis, lung, heart, liver, spleen and CNS. PCR products were separated on an agarose gel and imaged on a BioRad gel doc system.

Antipsychotic drug effects on MBNL1 expression

Levels of β-actin did not vary in the cortex of rats that received vehicle or chlorpromazine at different doses (F = 1.08, df = 3,19, p = 0.38). Moreover, levels of MBNL1 mRNA normalized to β-actin did not differ significantly in the cortex of rats treated with vehicle or chlorpromazine (F = 0.86, df = 3,19, p = 0.48; MBNL1 mRNA (mean±SEM), vehicle = 0.06±0.008 vs chlorpromazine 0.01 mg kg⁻¹ day ⁻¹ =0.07±0.007 vs chlorpromazine 0.1 mg kg⁻¹ day ⁻¹=0.07±0.007 vs chlorpromazine 1.0 mg kg⁻¹ day ⁻¹=0.04±0.009 vs chlorpromazine 10 mg kg⁻¹ day ⁻¹=0.06±0.012 MBNL1/β-actin ratio).

MBNL protein

There was no significant difference in the levels of MBNL protein in BA 46 from subjects with schizophrenia compared to controls (MBNL protein (mean±SEM): schizophrenia = 1.04±0.05 vs controls = 1.04±0.06 ratio IC; p = 0.99) measured using western blot analyses. Moreover, two-way ANOVA showed that there was no significant variance in MBNL levels with diagnosis (F < 0.001, df = 1,56, p = 0.99) or with DOI (F = 0.55, df = 1,56, p = 0.46) and there was no interaction between the variables (F = 0.07, df = 1,56, p = 0.78; MBNL protein (mean±SEM): SDOI schizophrenia = 1.06±0.07 vs controls = 1.08±0.09 ratio IC and LDOI schizophrenia = 1.020.07 vs controls = 1.000.09 ratio IC).

Discussion

We have utilized RFDD to measure the relative expression of approximately 12 500 mRNAs in BA 46 obtained after death from subjects with schizophrenia and in tissue from control subjects. Of these genes, 63 were shown to be differentially expressed in subjects with schizophrenia. This study suggests that genes involved in cell adhesion, signal transduction, development, transcriptional regulation, energy/metabolism and DNA/RNA binding proteins as well as genes encoding receptors, membrane transporters and structure/matrix proteins may be important in the pathology of schizophrenia.

Any comparison between our studies and those completed with microarrays needs to acknowledge the potential impact of methodological differences, differences in CNS regions studied and potentially different genetic backgrounds of the cases studied. However, even with the potential impact of these variables, our study supports previous findings that signal transduction [6, 7], energy/metabolism [8–10] and receptors [6, 12, 13] are all affected in schizophrenia. In addition, we did identify a key myelination gene as having altered expression levels in schizophrenia, supporting the notion that the disorder may involve abnormalities of myelination [14]. Our studies did not identify mRNA from genes involved in presynaptic function [5] or genes in the apolipoprotein family [5] as being abnormally expressed in schizophrenia but such discrepancies could be due to these genes not being amplified by the adaptors used in RFDD. In addition, our data highlight cell adhesion, development, control of structure/matrix and transcription as being particularly affected by the pathology of the disorder. Thus overall, our study using an open mRNA screening system has consolidated existing findings from studies using microarrays but has also expanded knowledge on key CNS systems that might be involved in the pathology of schizophrenia.

Studies on MBNL1

We focused our initial validation studies on the MBNL1 gene partly because of the availability of an antibody that allowed protein measurements. Our initial studies using real-time PCR confirmed that levels of MBNL1 mRNA were increased in BA 46 from subjects with schizophrenia. In addition they showed that this increase had both diagnostic (not different in BPD1) and regional (not altered in BA 9) selectivity. Of equal interest was the finding that the increased level of MBNL1 mRNA was subsequently detectable only in BA 46 from SDOI subjects. While it cannot be totally excluded that these differences are due to the duration of exposure to antipsychotic drugs per se, it would seem that altered levels of MBNL1 mRNA in SDOI schizophrenia is not simply due to exposure to antipsychotic because (i) MBNL1 mRNA was not altered in LDOI subjects or BD subjects who have been exposed to similar levels of antipsychotic drugs; (ii) levels of MBNL1 mRNA did not correlate with the FRADD in SDOI subjects; (iii) there was regional specificity (an outcome difficult to explain as a reaction to antipsychotic drug treatment); and (iv) it was not altered in the cortex of rats treated with an antipsychotic drug. Thus, overall, our finding suggests that the differential expression of genes in the CNS of subjects with schizophrenia may vary with the progression of the illness. Whether such changes in gene expression may in some way correlate with the changing symptom profile, which is suggested to occur with disease progression [25–29], is an important question that needs further investigation.

MBNL1 is one of three muscleblind proteins in the mammalian genome [30], all of which bind to double stranded CUG (dsCUG) RNA. The MBNL1 gene, first identified in Drosophila, was initially shown to be critical in photoreceptor differentiation [31] and subsequently in muscle development [32]. In studies of mammalian cells, the expression of MBNL1 has been shown to be increased in response to growth hormone in a murine 3T3-F442A pre-adipocyte/adipocyte differentiation model [33] and in follicular lymphoma biopsies in which the parent tumour had subsequently differentiated into an aggressive lymphoma [34]. Both these findings are consistent with mammalian MBNL1 playing a role in cellular differentiation. Extrapolating from data on mammalian MBNL1, increased levels of RNA from the MBNL1 gene in BA 46 from subjects with schizophrenia could be a marker of abnormal cellular differentiation or development, either outcome could cause the symptoms associated with the disorder. Significantly, abnormal expression of MBNL1 occurs in myotonic dystrophy [30]. This raises the possibility that abnormalities in muscle function that have been reported in subjects with schizophrenia [35] could be an outcome of abnormal expression of MBNL1 in the periphery. More pertinent to our findings is that subjects with myotonic dystrophy have significant cognitive impairment [36]. This suggests that abnormal expression of MBNL1 in the CNS from subjects with schizophrenia could be involved in the cognitive deficits associated with the disorder [37].

Our studies failed to show a change in MBNL protein in BA 46 from subjects with schizophrenia, making interpretation of the mRNA more difficult. Simplistically, the protein data could be interpreted as showing that changes in MBNL1 expression have not impacted on levels of encoded protein and hence MBNL function. However, an increasing understanding of the complexities involved in the generation of protein from genes [38] and the ability of RNA to modulate cellular function [39] makes such a conclusion both dated and presumptive. A more comprehensive understanding of the possible impact of overexpressing MBNL1 gene will be required to understand what role that gene may play in schizophrenia, particularly because MBNL1 has recently been shown to be important in the processes of regulating alternative splicing [40], a process critical in regulating normal protein expression [41].

Studies on DISC1

Both RFDD and real-time PCR showed that DISC1 mRNA was not altered in BA 46 from any diagnostic cohort. These findings support a previous study that showed that levels of DISC1 mRNA is not altered in the DLPFC or hippocampus from subjects with schizophrenia [42]. Because antipsychotic drug treatment in mice has complex effects on DISC1 expression [43] it is possible, but unlikely, that the absence of changed DISC1 mRNA in the CNS of subjects with schizophrenia and BPD1 could be due to antipsychotic drug treatment reversing changes in DISC1 expression in drug-naïve individuals. Importantly, the DISC1 gene has been proposed as a candidate gene for schizophrenia [44]. The notion that the DISC1 gene could be a candidate gene for schizophrenia arose from the discovery of a gene translocation in a Scottish family that affected both the DISC1 and DISC2 gene and that the presence of this translocation increased the risk of a family member having a psychiatric illness [24]. Subsequently, a number of studies have shown an association between this chromosomal abnormality and an increased risk of psychiatric illness, two of these studies being focused on the Scottish population as a whole [45, 46]. Other findings supporting DISC1 as a candidate gene for schizophrenia are the demonstration of a linkage between DISC1 and schizophrenia in a Finnish population [47] and haplotype association between the DISC1 gene and schizophrenia in a Taiwanese population [48].

The most obvious mechanism by which a gene mutation might cause some of the symptoms of schizophrenia would be by changing expression levels of the protein encoded by that gene. The present study would suggest that changes in DISC1 expression do not contribute significantly to changes in the molecular cytoarchitecture of the DLPFC in schizophrenia. However, the absence of a change in apparent DISC1 gene expression could be due to a number of factors. For example, there was no apparent association between the DISC1 gene and schizophrenia in a polygenetic population [49], which is more likely to be reflective of the polygenetic background of the cases used in our RFDD study. It has also been suggested that the translocation of DISC1 does not directly affect the expression of that gene but does affect the expression of other genes affected by the translocation [42], an hypothesis in line with the data on DISC1 in this study.

Studies on PCDH17

Unlike the agreement obtained between RFDD and real-time PCR on the expression of MBNL1 and DISC1, RFDD suggested that levels PCDH17 mRNA was decreased in schizophrenia but real-time PCR showed levels of mRNA to be increased in BA 46 and not changed in BA 9 from subjects with the disorder. These data show that, as with microarrays [6, 50], not all findings using high-throughput screening are readily translatable to other technologies. However, in the present study it is important to note that PCDH17 mRNA was first amplified in the RFDD methodology with non-specific primers whereas the real-time PCR primers were designed to amplify all PCDH17 mRNA. Therefore, because the PCDH17 gene is expressed in multiple splice forms [51], our data could indicate that there are differential changes in expression of the spliced variants of the PCDH17 gene in schizophrenia.

PCDH17 is a member of a family of approximately 80 proteins; the protocadherins being the largest group in the cadherin superfamily [52]. The exact role of PCDH17, formally PCDH 68, is yet to be elucidated [53]. However, the protocadherin family are known to be involved in the segregation of neuronal precursor populations, axonal outgrowth and synaptic formation [53] and clearly alterations in any of these functions could contribute to the onset of schizophrenia. Our study is focused on measuring levels of gene expression in the mature human CNS, where it is most likely the protocadherins are involved in synaptic formation, and therefore our findings would support the hypothesis that schizophrenia is a disorder of the synapse [4, 54].

In conclusion, the use of RFDD has identified changes in gene expression in BA 46 from subjects with schizophrenia and hence reinforced the argument that screening of the transcriptome in schizophrenia will identify new candidate markers for the disorder [55]. In addition, our validation studies have shown that understanding the role of such new candidate markers in schizophrenia is a major challenge that follows the creation of transcriptomic data sets for schizophrenia. Finally, a significant outcome from our early validation studies is that gene expression may change significantly with advancing duration of illness, as typified by our data on MBNL1 and PCDH17 in tissue from SDOI and LDOI subjects. This hypothesis deserves further exploration because it suggests that a critical period for changes in gene expression may be early after the onset of the disorder.

Footnotes

Acknowledgements

BD is an NHMRC Senior Research Fellow (400016). ES is the Ronald Phillip Griffith Research Fellow at the University of Melbourne and NARSAD Young Investigator. This research was supported by NHMRC Project Grant 192399, NIH R01 grant MH069691-01A1 and a Stanley Research Centre Grant.

References

1. Riley

McGuffin

. Linkage and associated studies of schizophrenia. Am J Med Genet 2000; 97: 23–44.

2. Tsuang

. Schizophrenia: genes and environment. Biol Psychiatry 2000; 47: 210–220.

3. American Psychiatric Association . Diagnostic and statistical manual of mental disorders4th edn. American Psychiatric Association, Washington, DC 1994.

4. Dean

. Signal transmission, rather than reception, is the underlying neurochemical abnormality in schizophrenia. Aust N Z J Psychiatry 2000; 34: 560–569.

5. Mirnics

Middleton

Marquez

Lewis

Levitt

. Molecular characterization of schizophrenia viewed by microarray analysis of gene expression in prefrontal cortex. Neuron 2000; 28: 53–67.

6. Hemby

Ginsberg

Brunk

Arnold

Trojanowski

Eberwine

. Gene expression profile for schizophrenia: discrete neuron transcription patterns in the entorhinal cortex. Arch Gen Psychiatry 2002; 59: 631–640.

7. Mirnics

Middleton

Stanwood

Lewis

Levitt

. Disease-specific changes in regulator of G-protein signaling 4 (RGS4) expression in schizophrenia. Mol Psychiatry 2001; 6: 293–301.

8. Middleton

Mirnics

Pierri

Lewis

Levitt

. Gene expression profiling reveals alterations of specific metabolic pathways in schizophrenia. J Neurosci 2002; 22: 2718–2729.

9. Iwamoto

Bundo

Kato

. Altered expression of mitochondria-related genes in postmortem brains of patients with bipolar disorder or schizophrenia, as revealed by large-scale DNA microarray analysis. Hum Mol Genet 2005; 14: 241–253.

10.

10. Altar

Jurata

Charles

. Deficient hippocampal neuron expression of proteasome, ubiquitin, and mitochondrial genes in multiple schizophrenia cohorts. Biol Psychiatry 2005; 58: 85–96.

11.

11. Hashimoto

Volk

Eggan

. Gene expression deficits in a subclass of GABA neurons in the prefrontal cortex of subjects with schizophrenia. J Neurosci 2003; 23: 6315–6326.

12.

12. Vawter

Crook

Hyde

. Microarray analysis of gene expression in the prefrontal cortex in schizophrenia: a preliminary study. Schizophr Res 2002; 58: 11–20.

13.

13. Mexal

Frank

Berger

. Differential modulation of gene expression in the NMDA postsynaptic density of schizophrenic and control smokers. Brain Res Mol Brain Res 2005; 139: 317–332.

14.

14. Hakak

Walker

. Genome-wide expression analysis reveals dysregulation of myelination-related genes in chronic schizophrenia. Proc Natl Acad Sci USA 2001; 98: 4746–4751.

15.

15. Vawter

Barrett

Cheadle

. Application of cDNA microarrays to examine gene expression differences in schizophrenia. Brain Res Bull 2001; 55: 641–650.

16.

16. Mimmack

Ryan

Baba

. Gene expression analysis in schizophrenia: reproducible up-regulation of several members of the apolipoprotein L family located in a high-susceptibility locus for schizophrenia on chromosome 22. Proc Natl Acad Sci USA 2002; 99: 4680–4685.

17.

17. Dean

Keriakous

Thomas

Scarr

. Understanding the pathology of schizophrenia: the impact of high-throughput screening of the genome and proteome in postmortem CNS. Curr Psychiatry Rev 2005; 1: 1–9.

18.

18. Sutcliffe

. Open-system approaches to gene expression in the CNS. J Neurosci 2001; 21: 8306–8309.

19.

19. Weinberger

. Schizophrenia and the frontal lobe. Trends Neurosci 1988; 11: 367–370.

20.

20. Hill

Keks

Roberts

. Problem of diagnosis in postmortem brain studies of schizophrenia. Am J Psychiatry 1996; 153: 533–537.

21.

21. Kingsbury

Foster

Nisbet

. Tissue pH as an indicator of mRNA preservation in human post-mortem brain. Brain Res Mol Brain Res 1995; 28: 311–318.

22.

22. Torrey

Webster

Knable

Johnston

Yolken

. The Stanley foundation brain collection and neuropathology consortium. Schizophr Res 2000; 44: 151–155.

23.

23. Kane

McGlashan

. Treatment of schizophrenia. Lancet 1995; 346: 820–825.

24.

24. Millar

Wilson-Annan

Anderson

. Disruption of two novel genes by a translocation co-segregating with schizophrenia. Hum Mol Genet 2000; 9: 1415–1423.

25.

25. Marengo

Harrow

Herbener

Sands

. A prospective longitudinal 10-year study of schizophrenia's three major factors and depression. Psychiatry Res 2000; 97: 61–77.

26.

26. Meagher

Quinn

Bourke

. Longitudinal assessment of psychopathological domains over late-stage schizophrenia in relation to duration of initially untreated psychosis: 3-year prospective study in a long-term inpatient population. Psychiatry Res 2004; 126: 217–227.

27.

27. Ho

Andreasen

Nopoulos

Arndt

Magnotta

Flaum

. Progressive structural brain abnormalities and their relationship to clinical outcome: a longitudinal magnetic resonance imaging study early in schizophrenia. Arch Gen Psychiatry 2003; 60: 585–594.

28.

28. Bustillo

Lauriello

Rowland

. Longitudinal follow-up of neurochemical changes during the first year of antipsychotic treatment in schizophrenia patients with minimal previous medication exposure. Schizophr Res 2002; 58: 313–321.

29.

29. Mathalon

Sullivan

Lim

Pfefferbaum

. Progressive brain volume changes and the clinical course of schizophrenia in men: a longitudinal magnetic resonance imaging study. Arch Gen Psychiatry 2001; 58: 148–157.

30.

30. Kanadia

Johnstone

Mankodi

. A muscleblind knockout model for myotonic dystrophy. Science 2003; 302: 1978–1980.

31.

31. Begemann

Paricio

Artero

Kiss

Perez-Alonso

Mlodzik

. muscleblind, a gene required for photoreceptor differentiation in Drosophila, encodes novel nuclear Cys3His-type zinc-finger-containing proteins. Development 1997; 124: 4321–4331.

32.

32. Artero

Prokop

Paricio

. The muscleblind gene participates in the organization of Z-bands and epidermal attachments of Drosophila muscles and is regulated by Dmef2. Dev Biol 1998; 195: 131–143.

33.

33. Shang

Thompson

Teasdale

Brown

Waters

. Genes induced by growth hormone in a model of adipogenic differentiation. Mol Cell Endocrinol 2002; 189: 213–219.

34.

34. Akasaka

Lossos

Levy

. BCL6 gene translocation in follicular lymphoma: a harbinger of eventual transformation to diffuse aggressive lymphoma. Blood 2003; 102: 1443–1448.

35.

35. Meltzer

Jackman

Arora

. Brain and skeletal muscle monoamine oxidase activity in schizophrenia. Schizophr Bull 1980; 6: 208–212.

36.

36. Perini

Menegazzo

Ermani

. Cognitive impairment and (CTG)n expansion in myotonic dystrophy patients. Biol Psychiatry 1999; 46: 425–431.

37.

37. Meltzer

. Cognitive factors in schizophrenia: causes, impact, and treatment. CNS Spectrum 2004; 9: 15–24.

38.

38. Sharma

. Schizophrenia, epigenetics and ligand-activated nuclear receptors: a framework for chromatin therapeutics. Schizophr Res 2005; 72: 79–90.

39.

39. Meyer

Temme

Wahle

. Messenger RNA turnover in eukaryotes: pathways and enzymes. Crit Rev Biochem Mol Biol 2004; 39: 197–216.

40.

40. Ho

Charlet

Poulos

Singh

Swanson

Cooper

. Muscleblind proteins regulate alternative splicing. EMBO J 2004; 23: 3103–3112.

41.

41. Lewis

Green

Brenner

. Evidence for the widespread coupling of alternative splicing and nonsense-mediated mRNA decay in humans. Proc Natl Acad Sci USA 2003; 100: 189–192.

42.

42. Lipska

Peters

Hyde

. Expression of DISC1 binding partners is reduced in schizophrenia and associated with DISC1 SNPs. Hum Mol Genet 2006; 15: 1245–1258.

43.

43. Chiba

Hashimoto

Hattori

. Effect of antipsychotic drugs on DISC1 and dysbindin expression in mouse frontal cortex and hippocampus. J Neural Transm 2006; 113: 1337–1346.

44.

44. Ross

Margolis

Reading

Pletnikov

Coyle

. Neurobiology of schizophrenia. Neuron 2006; 52: 139–153.

45.

45. Thomson

Wray

Millar

. Association between the TRAX/DISC locus and both bipolar disorder and schizophrenia in the Scottish population. Mol Psychiatry: 657–668 2005; 10: 616.

46.

46. Zhang

Sarginson

Crombie

Walker

St Clair

Shaw

. Genetic association between schizophrenia and the DISC1 gene in the Scottish population. Am J Med Genet B Neuropsychiatry Genet 2006; 141: 155–159.

47.

47. Ekelund

Hennah

Hiekkalinna

. Replication of 1q42 linkage in Finnish schizophrenia pedigrees. Mol Psychiatry 2004; 9: 1037–1041.

48.

48. Liu

Fann

Liu

. A single nucleotide polymorphism fine mapping study of chromosome 1q42.1 reveals the vulnerability genes for schizophrenia, GNPAT and DISC1: association with impairment of sustained attention. Biol Psychiatry 2006; 60: 554–562.

49.

49. Devon

Anderson

Teague

. Identification of polymorphisms within Disrupted in Schizophrenia 1 and Disrupted in Schizophrenia 2, and an investigation of their association with schizophrenia and bipolar affective disorder. Psychiatry Genet 2001; 11: 71–78.

50.

50. Tkachev

Mimmack

Ryan

. Oligodendrocyte dysfunction in schizophrenia and bipolar disorder. Lancet 2003; 362: 798–805.

51.

51. Redies

Vanhalst

Roy

. delta-Protocadherins: unique structures and functions. Cell Mol Life Sci 2005; 62: 2840–2852.

52.

52. Junghans

Haas

Kemler

. Mammalian cadherins and protocadherins: about cell death, synapses and processing. Curr Opin Cell Biol 2005; 17: 446–452.

53.

53. Frank

Kemler

. Protocadherins. Curr Opin Cell Biol 2002; 14: 557–562.

54.

54. Owen

O'Donovan

Harrison

. Schizophrenia: a genetic disorder of the synapse?. BMJ 2005; 330: 158–159.

55.

55. Bahn

Augood

Ryan

Standaert

Starkey

Emson

. Gene expression profiling in the post-mortem human brain: no cause for dismay. J Chem Neuroanat 2001; 22: 79–94.