< span>Nancy L. Johnston, Andrew D. Shore, Robert H. Yolken, E. Fuller Torrey, and the Stanley Neuropathology Consortium. Stanley Neurovirology Laboratory of the Johns Hopkins University, Baltimore, MD; Neurosciences Center at St. Elizabeth’s, Washington DC.

< span>Two-dimensional gel electrophoresis is a means to compare protein species and levels between samples. It is usually done among tightly controlled samples obtained from cells or experimental animals where the variables can be limited to answer an experimental question. In order to study human postmortem samples where there is little control over these variables, we undertook a study of a set of 96 samples prepared from human frontal cortices of healthy and mentally ill individuals. For the most part, protein levels were remarkably similar between individuals. We compared variations in 217 protein levels against known demographic factors such as age, PMI, pH (an indicator of agonal state) freezer storage interval, alcohol and drug abuse, body mass index, brain, weight, suicide, and mental diagnosis. There were 60 proteins that were significantly affected (p<0.02) by these factors; 50 were affected by a single variation, 8 were affected by two and 2 were affected by three factors. Factors which most influenced protein levels were freezer storage interval (N=11), premortem factors (pH) (N=10), death by carbon monoxide poisoning (N=9), and suicide other than carbon monoxide poisoning (N=9). We identified several protein species that were substantially lower in the control categories than in one or more of the mentally ill categories. Of these latter proteins, five were chosen for sequencing. None of the candidates for sequencing varied with age, body mass, brain weight, freezer storage interval, PMI or pH.

< span>


< span>Frances Yee, Nancy Johnston, Flora Leister, Shuojia Li, Christopher Ross, E. Fuller Torrey, Robert Yolken and the Stanley Neuropathology Consortium. Stanley Neurovirology Laboratory and Depts. of Psychiatry and Neurosciences of the Johns Hopkins University, Baltimore, MD.

< span>We have investigated the hypothesis that the expression of brain RNAs may be altered in neuropsychiatric illnesses, such as schizophrenia and bipolar disorder. A modified version of the arbitrarily primed-polymerase chain reaction (AP-PCR) method (Welsh et al, Nucl. Acids Res. 20:4965, 1992) was used to determine whether RNAs are differentially expressed in cortical regions obtained postmortem from individuals with schizophrenia and unaffected individuals with no previous diagnosis of mental illness.

< span>Several RNA transcripts had higher expression levels in the frontal cortex of an individual with schizophrenia compared to the normal control. One of the candidate RNAs displayed 77% amino acid homology to the Pol polyprotein of simian retrovirus (SRV-1), which is a primate type-D retrovirus. A second transcript was nearly identical to the BD1 zinc finger protein (>99% predicted amino acid similarity), which interacts with the tax protein of human T-cell leukemia virus to increase interleukin-3 transcription. The third candidate RNA was similar (64% nucleotide homology) to a simian paramyxovirus (SV-5). A fourth transcript encodes a putative peptide distantly related to an envelope protein of an avian retrovirus, avian Reticulosis virus.

< span>These candidate clones were then screened for disease specificity using (oligo dT-primed) semi-quantitative reverse transcription-PCR on frontal cortical RNAs extracted postmortem from individuals with schizophrenia (n=17), bipolar disorder (n=17), non-psychotic depression (n=12), Huntington’s Disease (n=8) and controls without history of psychiatric illness (n=15). These reactions were duplexed with a second PCR using primers for glyceraldehyde-3-phosphate dehydrogenase, which is a housekeeping gene and serves as an internal control for normalizing RNA content across samples. Preliminary data indicate that ther are significant differences in the levels of the simian retrovirus-like transcript in the bipolar group (p<0.004 by ANOVA). Our preliminary findings indicate that viral and virus-associated RNA transcripts are expressed in the brains of some individuals with schizophrenia and bipolar disorder, and may play an important role in disease pathogenesis.

< span>


< span>Yeping Sun, Robert H. Yolken and The Stanley Neuropathology Consortium. Stanley Neurovirology Laboratory of the Johns Hopkins University, Baltimore, MD.

< span>Gene expression of the frontal cortex tissue from schizophrenia and bipolar disorder were analyzed by SAGE. Briefly, brain RNA were extracted and converted to full-length cDNA by Capfinder cDNA synthesis kit. The resulting cDNA was then fragmented to 9-base pair tags by the action of a restriction endonuclease. NIaIII and a type II restriction enzyme, BsmFI, so that each tag represents a copy of the gene transcript. We have analyzed a total of 6,000 tags from schizophrenia, bipolar and normal controls. We found a number of genes were expressed differentially in affected brains as compared to normal controls. Among them, the most prominent difference found for schizophrenia was increased expression of mitochondrial mRNA for cytochrome oxidase and mitochondrial atpase. Others included serotonin transportor gene and human biliary glycoprotein gene were expressed in a significantly higher level in both schizophrenia and bipolar disorder but not in normal brain. Transcripts of RACH1 gene and XB gene are expressed in a significantly higher level in bipolar disorder only. Ongoing studies are directed to verify these differences by polymerase chain reaction and Northern hybridization assay in a larger number of samples.


< span>Indra Dé, P. Allen, CA Ross, RH Yolken, EF Torrey and the Stanley Neuropathology Consortium. The Stanley Neurovirology Laboratory and the Depts. of Psychiatry and Neuroscience of the Johns Hopkins University, Baltimore, MD; Laboratory of Molecular and Cellular Neuroscience, The Rockefeller University, New York, NY; NIMH Neurosciences Center at St. Elizabeth’s, Washington DC

< span>Spinophilin, localizes to the head and neck region of dendritic spines, and binds to protein phosphatase 1 (PP1), which regulates neurotranscmitter receptors, ion channel function, and the actin microfilament network. Thus spinophilin may serve to localize and modulate PP1 activity in neurons by presenting the enzyme to targets within the dendritic spine compartment, thereby regulating synaptic function and plasticity. Such properties make spinophilin a candidate postsynaptic marker in the study of complex neuropsychiatric disorders like schizophrenia, where alteration in synaptic function and/or plasticity may be one of many factors involved in the pathophysiology of the disease. We therefore sought to determine whether spinophilin levels are altered in postmortem brains from schizophrenics.

< span>A preliminary screening of 24 protein extracts of postmortem brains from individuals with schizophrenia (13) and normal controls (11) by immunoblotting, using an antispinophilin antibody was undertaken. Qualitatively, the majority of the brains tested showed adequate amounts of spinophilin. Qualitative inspection suggested less reactivity in the brains from schizophrenics. These results prompted a more controlled study of spinophilin expression in a larger sample set: 30 postmortem brains from individuals with schizophrenia (15) and matched normal controls (15) using the constitutive heat shock protein (HSC) 70 as control. The results of this study will be presented and its implications in the evolving pathophysiology of schizophrenia will be discussed.

< span>


< span>Håkan Karlsson, Flora Leister, Nancy L. Johnston, Robert H. Yolken and the Stanley Neuropathology Consortium. Stanley Neurovirology Laboratory of the Johns Hopkins University, Baltimore, MD.

< span>Several factors suggest a possible viral involvement in the pathogenesis of schizophrenia. The DNA-viruses of the herpes family has received much attention due to their neurotropism. The aim of the present study was to screen CSF’s obtained postmortem from affected patients and unaffected controls for the presence of members of the herpesvirus family.

< span>DNA was extracted from cerebrospinal fluids obtained postmortem from 17 schizophrenic individuals, 12 individuals with bipolar disorder, 12 control individuals with other mental illnesses and 8 unaffected individuals. These samples were examined for the presence of nucleic acid sequences from viruses of the herpes virus family. Nested polymerase chain reactions (PCR) were used to detect the following members of this family; herpes simplex virus (HSV)-1 and -2, human herpesvirus (HHV) -6 and -8, cytomegalovirus (CMV) and Epstein-Barr virus (EBV).

< span>HSV-1, -2, HHV-6, CMV or EBV was not detected in any of the CSF’s in this study. Testing for HHV-8 resulted in several products, the specificity of which needs to be confirmed by Southern blotting. These results are in agreement with other studies. This does, however, not rule out the possibility that members of this family of viruses are present in the brains of the affected individuals, although the viral sequences could not be detected in the CSF. Viruses, not identical to prototype herpesviruses, may also be present in the CSF and/or brains of these patients.

< span>We are currently using a pan-retrovirus PCR detection system for screening the same CSF’s as well as tissue from the frontal cortex of brains from affected individuals and normal controls for the presence of RNA sequences of retroviral origin. Preliminary data indicate that retroviral RNA is not present in the ventricular fluids in any of these samples. However, screening frontal cortexes for the presence of retroviral RNA’s gave rise to bands of various size in many samples. The identity of these PCR products is currently being investigated.

< span>


< span>Raphael Viscidi, Elizabeth Rubalcaba, Mario Solari, and Robert Yolken. Department of Pediatric Infectious Diseases, Johns Hopkins University, Baltimore, MD.

< span>The human genome contains diverse endogenous retrovirus-like elements. In order to screen human populations for disease-associated retroviral sequences, a generic detection method for this class of sequences would be desirable. We have developed a PCR method for the differential display of retroviral-like sequences in the human genome that exploits characteristic features of retroviral replication and transcription. The method makes use of the primer binding site for a host tRNA, which is adjacent to the 3′ end of the LTR and serves to initiate reverse transcription, and an upstream primer consisting of a DNA binding motif for transcription factors commonly found in retroviruses. For the current studies four tRNA primers [tRNAHis(HIS), tRNALys1 (LYS1), tRNALys2(LYS2) and tRNAPro(PRO)] were used in combination with primers for four DNA response elements [cAMP response element (CRE), CCAAT/enhancer binding protein recognition element (C/EBP), glucocorticoid response element (GRE) and an ets-responsive region (ERR)]. The PCR products were resolved on a nondenaturing polyacylamide gel. DNA bands of interest were excised from the gel, re-amplified by PCR, cloned and sequenced. The method was applied to the analysis of genomic DNA from 9 monozygotic twins discordant for schizophrenia and 7 normal twins. Band patterns were scored for qualitative and/or quantitative differences between individuals within a discordant pair or for differences between affected and normal twins. Difference-bands detected in one or more subjects using the indicated primer pairs included: 600bp PRO-CRE band (2 subjects), 125 bp HIS-ERR band (3 subjects), 175 bp HIS-GRE band (3 subjects), 325 bp HIS-GRE band (1 subject) and 400 bp PRO-ERR band (2 subjects). The sequences of 3 out of 11 groups of clones from the 600bp PRO-CRE band (PRO-CRE clones A, B and C) and 2 of 10 clones from the 125 bp HIS-ERR band showed significant homology to retroviral LTR sequences in a BLAST search. PCR of genomic DNA from the 9 discordant twins and 7 normal twins, using specific primers for the sequence of PRO-CRE clone A, B, and C showed bands of the expected size and of equivalent intensity in all subjects for each of the sequences. Analysis of clones from other bands and retesting of subjects with sequence specific primers is in progress. A simple and efficient method is described for detection of defined classes of retroviral sequences in the human genome.

< span>


< span>Husseini K. Manji, Joseph M. Bebchuk, and Guang Chen. Molecular Pathophysiology Program, WSU School of Medicine

< span>Bipolar disorder is a common and chronic illness with significant morbidity and mortality, and there is a clear need to develop new agents which work quickly, potently, specifically, and with few side effects. To identify the biochemical mechanisms in the brain by which mood-stabilizing agents bring about their therapeutic effects, we have been actively investigating the PKC signaling pathway. PKC represents a family of isozymes which are highly enriched in brain, and play a major role in regulating multiple aspects of neurotransmission and long term neuroplastic events, including gene expression. We have found that chronic, but not acute, lithium administration to rats produces a significant decrease in [3H]PDBU binding in several limbic and limbic-related structures, most notably subiculum and CA1 region. Immunolabeling using specific antibodies have revealed isozyme-specific decrease in levels of membrane-associated PCK a and e, findings which parallel those that we and others have also observed in cell culture. We have also investigated the effects of another effective antimanic and mood-stabilizing agent, valproate (VPA) on the PKC signaling system. We have found that chronic, but not acute, VPA also produces strikingly similar reductions in PKC activity and isozyme-selective reductions in the levels of PKC a and e. Since PKC plays a major role in regulating neuronal excitability and neurotransmitter release, we have postulated that PKC inhibition may play a role in the antimanic effects of these agents. To investigate this more directly, we have initiated a study investigating the efficacy of Tamoxifen (at doses sufficient to inhibit PKC) in the treatment of acute mania; preliminary results in a small group of manic patients suggests that Tomoxifen possesses antimanic efficacy. Together, these results suggest that PKC-modulators may represent a truly novel class of drugs for the treatment of Bipolar Disorder.

Supported by the Theodore and Vada Stanley Foundation.


< span>Kathleen A. Hay and Richard B. Tenser. Penn State University College of Medicine, Hershey, PA.

< span>Herpes simplex virus (HSV) latent infection of peripheral nervous system sensory and autonomic neurons has been extensively described in humans and experimental animals. Latent infection of the central nervous system (CNS), however, has been only sporadically noted, usually as an observation of occasional cells showing evidence of in situ hybridization positive signal. In the present study we report results of HVS latent infection of CNS with reproducible in situ hybridization detection of viral RNA, and also polymerase chain reaction (PCR) detection of viral DNA. Balb/c mice were nasally infected with HSV-type 1 (20 ul, 105 PFU) by usual methods. During the period of latency associated transcript (LAT) detection by in situ hybridization using a radiolabeled oligonucleotide probe (J. Virol. 1996, 20:1271). Alternately, CNS tissue was investigated using PCR methods to detect HSV DNA (Virology 1993;195:337). HSV LAT was detected in facial nucleus neurons in 9 of 12 mice (average 173 LAT-positive neurons). LAT was also present in several diencehalic periventricular cells in 3 of 12 mice. PCR results showed HSV DNA most prominently in brainstem. Results of this preliminary study demonstrated HSV latent infection of the CNS. It is planned to continue mapping areas of CNS latency, to determine altered function of latently infected cells and to investigate reactivation.

< span>


< span>Jeng-Yang Ling, Tsuey-Ming Chen and William G. Stroop. University of Arkansas for Medical Science

< span>Herpes simplex viruses, type 1 (HSV-1) and 2 (HSV-2), are double-stranded, enveloped DNA viruses and the HSV infection of central nervous system (CNS) can be very serious or fatal. Three intertypic recombinant viruses were generated between syn HSV-1 strain +GC and syn+ HSV-2 strain HG52p18. +GC causes seizures and high mortality whereas HG52p18 is non-virulent as well as weakly neuroinvasive in New Zealand white rabbits following intranasal inoculations. Recombinant R1 contains an approximate 6.4 kbp DNA fragment from +GC containing intact UL 27 and UL26/26.5 genes, plus portions of the UL25 and UL28 genes on the background of HG52p18. Two reciprocal recombinants were constructed which contained chimeric UL27 genes in which the majority of +GC DNA was replaced with the homologous DNA from the non syncytial HSV-2 HG52p18 strain. Restriction enzyme mapping analysis revealed that reciprocal recombinant R3 contained only the carboxyl ends of the UL27 and UL25 genes together with the intact UL26 and UL26.5 genes, and reciprocal recombinant R4 contained almost all of the UL27 gene but only part of the UL 26 and UL26.5 genes of HG52p18 on the background of +GC. In cell culture, both +GC and R1 induced syncytia, and HG52p18, R3 and R4 were non-syncytiagenic in vitro.

< span>Following intranasal inoculation of rabbits, R1 was significantly more virulent than HG52p18 and was as virulent as the +GC parental virus. R1 also induced seizures. Reciprocal recombinant viruses, R3 and R4, were totally nonvirulent in rabbits after intranasal inoculation. Histopathological analysis and in situ hybridization were performed to study the neuroinvasion and tropism of the parental viruses and recombinants for central nervous system tissue. Histopathological analysis of brain tissues from HG52p18- and R1-infected rabbits revealed that both viruses efficiently invaded the trigeminal system. R1 infected rabbits had histologic evidence of infection and virus specific nucleic acids in their higher order olfatory centers. R3 and R4 were less invasive in the central nervous system olfactory pathway than their parental virus, +GC

< span>


< span>JH Gilmore, LF Jarskog, H Xiao. Department of Psychiatry, University of North Carolina School of Medicine

< span>Exposure to infection during early brain development has been associated with increased risk of schizophrenia and other neurodevelopmental disorders. We have hypothesized that general factors related to an immune response to infection, especially cytokines, play a role in this association, as cytokines can regulate neuronal development. In an ongoing effort to develop an animal model to study the impact of the systemic immune response to infection on the developing brain, we examined the effect of infection on the mRNA expression of the neurotropic factors brain derived neurotropic factor (BDNF), neurotrophin-3 (NT-3) and nerve growth factor in adult rats. Adult rats (n=5 per condition) received intraperitoneal injections of 5mg/kg of e. coli lipopolysaccaride (LPS) or saline. 4 hours after injection, rats were sacrificed and cortex and hippocampus were dissected. Changes in mRNA expression were measured in a semi-quantitative way using RT-PCR with oligonucleotide standards (MIMICS). LP decreased BDNF mRNA in the cortex by approx. 22% (p=0.0006), and a trend toward an increase in NGF mRNA in cortex (approx. 10%, p=0.1); there were no changes in the hippocampus. Preliminary studies at day 14, (n=3), found a decrease in NT-3 mRNA (approx. 24%, p-0.03) and a decrease in NGF mRNA (approx. 14%, p-0.09) in the hippocampus; there were no significant changes in the cortex. These results indicate that acute exposure to LPS does alter CNS expression of neurotropic factor mRNA and supports the hypothesis that early exposure to systemic infection can have a significant impact on the development of the brain. In this model, the effect of infection on neurotropic factor mRNA is specific to type of neurotropic factor, brain structure, and developmental age.

< span>


< span>Emilia L. Oleszak, Brad Hoffman, Nile Harvey, Christos D. Katsetos. Fels Inst. for Cancer Research & Molecular Biology, Departments of Biochem. & Neurol., Dept. Microbiol & Immunol., Temple University School of Medicine, Philadelphia, PA.

< span>Theiler’s murine encephalomyelitis virus (TMEV) induces in susceptible strains of mice a biphasic disease. Early acute disease resembles polioencephalomyelitis and is followed by a chronic primary demyelinating disease. During early acute disease, the virus replicates predominantly in gray matter. After a few weeks, the virus migrates to the white matter of the spinal cord and persists at low levels in macrophages and glial cells. Resistant strains of mice develop only early acute disease. Both phases of the disease are associated with perivascular and meningeal infiltrates. The mechanism which leads to demyelination is unknown. The immunopathology of demyelination induced by TMEV is similar to that of multiple sclerosis. We have examined by TUNEL assay whether the cells in the CNS of TMEV-infected mice are undergoing apoptosis. During early acute disease (6 and 10 days p.i.) large numbers of cells were undergoing apoptosis in the brain of TMEV-infected sensitive and resistant strains of mice. These were primarily mononuclear cells and were located mainly in perivascular cuffs in areas of encephalitis. Double immunohistochemical staining using an anti-CD-2 mab and the TUNEL assay demonstrated that the apoptotic cells were T lymphocyte. Very few other cell types showed apoptotic features; these were neurons and macrophages. However, in spite of heavy perivascular and leptomeningeal infiltrates during late chronic demyelinating disease, only rare apoptotic mononuclear cells and rare putative oligodendrocytes were identified in TMEV-infected mice. These results demonstrate that during early acute disease infiltrating T cells are eliminated in an apoptotic manner. This process may be associated with the clearance of the virus. However, T cells infiltrating the spinal cord of TMEV-infected SJL mice with late chronic demyelinating disease, only rarely undergo programmed cell death. These T cells may play a role in the pathogenesis of late chronic demyelinating disease.

< span>


< span>Husseini K. Manji and Guang Chen. Molecular Pathophysiology Program, WSU School of Medicine

< span>Mood-stabilizing agents require chronic administration for therapeutic efficacy, a temporal profile which suggests the possibility of alterations at the genomic level. We have undertaken a serious of studies to investigate the effects of lithium and VPA on the AP-1 family of transcription factors, and have found they increase AP-1 DNA binding activity, and also increase the expression of an AP-1 driven reporter gene. We have also been using mRNA RT-PCR differential display as a screening technique to identify gene(s) which are common targets for both lithium and VPA. Chronic administration of both drugs modulate the expression of several genes in limbic and limbic-related areas of rat brain. One sequence of cDNA fragments whose expression is markedly increased by both lithium and VPA has homology to a transcription factor, polyomavirus enhancer-binding protein 2(PEBP2) ß and acute myelogenous leukemia 1 (AML1)ß. We have further found that the chronic administration of both lithium and VPA increases DNA binding activity of the ab complex of PEBP2 in rat frontal cortex and hippocampus, as well as in human SY5Y cells. These novel effects may play a role in the therapeutic effects of mood-stabilizing agents, since PEB2ß has been postulated to play an important role in neuronal development and the survival of neurons; the effects of lithium and VPA on PEBP2 and PEBP2 regulated genes are currently under investigation. Overall, the results clearly demonstrate that mood-stabilizing agents selectively modulate gene expression in the CNS; the identification of the functions of these genes offers the potential not only for improved therapeutics, but may also provide important clues about the etiology/pathophysiology of manic depressive illness.

Supported by the Theodore and Vada Stanley Foundation.


< span>Babaz Azad. SRA Technologies Inc., Falls Church, VA

< span>In any biological or disease paradigm, it is important to determine whether all of the functional relevant gene expressions have been identified. Human genome is estimated to encode between 80-120000 different genes. In a typical cell, nearly 505 of total mRNA mass is composed of a relatively small number of known abundant mRNA species. This subpopulation has a tendency to mask display of other genes in discovery methodologies. Therefore, the mRNA complexity of a cell, mostly of rare abundance (1/2000 to 1/7000 copies/cell), is yet to be fully discovered.

< span>We have developed a high throughput fluorescent methodology for differential analysis of nearly all gene expressions (extremely low, low, medium, and high abundance mRNA’s). This Gene Expression Discovery System (GEDSSM) has been applied to multiple human disease related paradigms for digitization of gene expression survey and analysis. Quantitative fingerprints of Expressed Sequence Tags (EST’s) can be functionally correlated with the identification of novel targets for Therapeutics and Diagnostics. Significant technical improvements to cDNA amplification/normalization and arbitrary primed PCR based pattern generation allow display of nearly all genes involved in a relevant paradigm or pathway. Full integration of multicolor fluorescent pattern characterization, EST isolation, sequencing, and microarray based screening is currently under development.

< span>


< span>Cassandra Smith, Natasha Broude, Sarah Malpel, Joel Graber, Michael Lewis. Boston University, Center for Advanced Biotechnology and Departments of Biomedical Engineering, Biology, and Pharmacology.

< span>Positional cloning approaches are guaranteed to work provided appropriate families are identified and enough resources are available to test a large number of genetic markers. Given that there are an estimated 75,000 human genes, it is quite clear that alternatives to conventional positioning cloning would be quite valuable. One alternative method is differential display (DD). DD monitors gene expression levels. In DD, cDNAs from different cells are displayed (fractioned) by size for comparison. Although DD has proven useful in isolating important genes in breast cancer and some other studies, DD has not been useful in other studies. This method is technically difficult and usually comparisons are made between random cDNAs expressed at high levels. We have developed a genomic differential display (GDD) and a targeted DD (TDD), methods which use interspersed repeat sequences to focus comparisons on targeted classes of genes such as those containing (CAG)n repeating sequences. Our method has been demonstrated to detect true differences between samples. The differences may be insertions, deletions, rearrangements as well as some single base differences. This method was used to isolate differences between monozygotic twins discordant for schizophrenia and to compare genomic DNA from different regions of the brain.


< span>N. Leigh Anderson, Christine Sims, and Jean-Paul Hofman. Large Scale Biology Corporation, Rockville, MD.

< span>Abundance of a series of more than 200 protein spots were compared in specimens of 100 human brains analyzed by 2-D electrophoresis. The samples represent individuals with schizophrenia, depression, bipolar disorder, Huntington’s, miscellaneous, or no psychiatric illness. No single protein appeared to unequivocally discriminate between the groups. Using a subset of these proteins selected as showing some potential association (P<0.01) with one of the diseases, principal components analysis (PCA) revealed a set of proteins that in combination appeared to distinguish unaffected individuals from those with depression. This result suggested that panels of molecular markers, particularly proteins, may be useful in uncovering the biochemical basis of minor psychiatric diseases. Recent developments in mass spectrometry are likely to allow identification of these proteins in the near future.

< span>

Poster Presentations


< span>Nancy L. Johnston, Robert H. Yolken, E. Fuller Torrey and the Stanley Neuropathology Consortium. Stanley Neurovirology Laboratory of the Johns Hopkins University School of Medicine, Baltimore, MD.

< span>An objective and comprehensive method of looking at differences between the brains of mentally ill and unaffected individuals is to use subtractive techniques to compare the RNA from the brains of cases and controls. Messages that are present in significantly different levels in the brains of affected individuals may be identified and subsequently screened against a bank of samples to verify their correlation with disease. We used this approach to compare RNA derived from the brains of schizophrenic, bipolar and unaffected individuals. Comparisons of frontal cortices and posterior hippocampus showed a number of differentially expressed messages. This was further verified by screening against 80 cDNAs derived from the frontal lobes of individuals of varying diagnoses. This approach provides a viable method to identify messages potentially involved in mental illness.


< span>Linda D. Bobo, Jose D. Paltan-Ortiz, Mary M. Herman, Nathaniel C. Briggs, Juraj Cervenak, Robert H. Yolken, Melvyn P. Heyes, and E. Fuller Torrey. Stanley Neurovirology Laboratory of the Johns Hopkins University School of Medicine, Baltimore, MD; Stanley Foundation Research Program, NIMH Neuroscience Center, Washington DC; NIMH Neurochemistry Laboratory, Section on Analytical Biochemistry, Clinical Science Laboratory, Bethesda, MD.

< span>Introduction: We hypothesized that elevated CNS IL-2 receptor alpha (IL-2R) and quinolinic acid (QA) levels could have resulted from liver pathology in the presence or absence of specific epidemiologic risk factors. Since abnormal liver function can alter blood-brain permeability and astrocytic metabolism, our goal was to determine if CNS IL-2R or QA elevations that were significantly associated with schizophrenia (SZ) or bipolar disorder (BPD) occurred in the presence of specific liver pathology.

< span>Methods: The postmortem population consisted of cases with either SZ, BPD, or unipolar depression (UPD), and normal controls (NC). General autopsy findings were determined by medical examiners. Demographic, and pre- and postmortem variables relating to disease severity, life style, medications, systemic disease, and manner of death were classified. Psychiatric diagnoses were determined using DSM-IV criteria after chart review by 2-3 psychiatrists. Formalin-fixed, paraffin-embedded sections of liver and brain were independently evaluated by two neuropathologists who were masked as to psychiatric category and epidemiologic variables. Histopathologic findings were graded from 0 (no pathology) to 3 (marked) depending on the severity of the changes. Liver observations included: 1)fatty metamorphosis; 2) fibrosis and/or cirrhosis; 3) acute or chronic inflammation; 4) bile duct proliferation; and 5) cytologic changes in hepatocytes. Brain sections (including silver stains) from the frontal, temporal and occipital lobes, and from the cerebellar vermis were examined for the following: 1) evidence of persistent gliosis or recent glial activation, including hepatic-associated Alzheimer type II astrocytic changes; 2) presence of edema or hemorrhages; 3) type and location of inflammatory infiltrates (meningeal, perivascular or parenchymal); 4) neuronal changes; and 5) arteriosclerosis. IL-2R levels were determined in pg/ml from cisternal fluid (CSF) using a sandwich ELISA. QA and [13C8] Quin were quantified as their dihexafluoroisopropanol esters by electron capture negative/mass spectrophotemetry in pmol/g from frontal lobe and cerebellar cortex tissue and from CSF in pm/ml.

< span>Preliminary Results:Quinolinic acid. There was a significant association of elevated QA levels in cerebellar tissue of SZ compared to NC (p=.001) and UPD (p=.002). Median QA levels of cerebellar and frontal tissue, respectively, were: 186 and 120 for SZ, 134 and 98 for BPD, 81 and 72 for UPD, and 94 and 83 for NC. In combined cerebellar and frontal regions, elevated QA levels were stratified by the presence of significant liver pathology, i.e. moderate to severe fatty metamorphosis with fibrosis or inflammation, were 68.8% (11/16) for SZ, 81.8% (9/11) for BPD, 60% (6/10) for UPD, and 61.9% (13/21) for NC. In contrast, low QA levels stratified by significant liver pathology were 13.3% (2/15) for SZ, 27.4% (3/11) for BPD, 18.2% (2/11) for UPD, and 12% (3/25) for NC. The data indicate that there was a strong association of elevated brain QA levels with significant liver pathology (p<.0001; OR=10.67, 95% Cl:4.14-18.27) for all groups. Elevated CSF QA levels were significantly associated with SZ as compared to NC (p=.008). There was a similar trend overall for elevated CSF QA levels to be associated with significant liver pathology as compared to low QA levels. By staining methods used, there was no common type or severity grade of brain histopathologic finding for cases which had elevated tissue or CSF QA levels for there respective diagnosis either in the presence or absence of liver pathology. IL-2R. There was a significant association of elevated CSF IL-2R levels with SZ (p=.02) and BPD (p=.009) compared to NC. Median CSF IL-2R levels were 81 for SZ, 80 for BPD, 58 for UPD, and 39 for NC. IL-2R elevations stratified by significant liver pathology were 50% (3/6) for SZ, 50% (2/4) for BPD, 40% (2/5) for UPD, and 33.3% (3/9) for NC. In contrast, low IL-2R level stratified by significant liver pathology were 40% (4/10) for SZ, 22.2% (2/9) for BPD, 33.3% (2/6) for UPD and 25% (2/8) for NC. The association of elevated Il-2R with liver pathology was not significant.

< span>Conclusions: The results suggest that pathology in the liver may in part contribute to the finding of elevated QA in cerebellar tissue. In order to better understand the association with schizophrenia, liver QA levels will be correlated with disease categories and CNS QA levels. In terms of CNS IL-2R elevations, a larger sample size is being evaluated in order to reliably interpret the association with significant liver pathology. Other ongoing studies include correlations of medication (including neuroleptics), alcohol and other substance abuse, agonal state, viral hepatitis status, presence of vasculitis, and postmortem factors with liver pathology and QA and IL-2R levels. In addition, since specific brain histopathology was not seen in association with elevated QA levels by the microscopic methods used, brain sections will be evaluated by immunohistochemistry for specific markers of astrocytic and microglial activation. Finally, use of this approach as a model for quality control of postmortem samples may elucidate why elevated CNS immuno-inflammatory markers were found in psychiatric subgroups when stratified by liver/brain pathology and epidemiologic variables.

Interleukin 2 Soluble Receptor Levels in Schizophrenia and Deliberate Self-harm (DSH)

< span>Fiona Gaughran, H.S. Keeley, P. Corcoran. Professional Psychiatric Unit, Cork University Hospital; National Suicide Research Foundation, Ireland

< span>Background: Elevated Soluble Interleukin 2 Receptor Levels (sIL2R) have been reported in both schizophrenia and in DSH. Multiple attempts at DSH occur in a hgh percentage of patients with schizophrenia. We examined all sIL2R levels in people after DSH and with acute schizophrenia to clarify the relationship between these conditions and sIL2R.

Methods: Persons admitted follwoing an episode of DSH were included as well as people admitted with acute schizophrenia or schizoaffective disorder. Psychiatric diagnoses were made using the Structured Clinical Interview for DSM IIIR or according to ICD 10 criteria. Serum was taken with informed consent and tested for sIL2R levels with a Quantokine ELISA kit.

Results: There was a striking increase in those who had attempted self-harm more than once versus controls (p<0.0001). There was no difference between single episode DHS patients and controls.

Patients with a diagnosis of schizophrenia had highe sIL2R than controls (p=0.0003). Of 36 patients with schizophrenia, 17 (47.2%) had never tried to harm themselves, 5 (13.9%) had moade one attempt and 14 (38.9%) had made one attempt and 14 (38.9%) multiple attempts. The patients with schizophrenia but no DSH had higher sIL2R levels than controls (medium = 1049, p=0.0051), but in those schizophrenics with a history of multiple DSH, sIL2R levels were greater again (P=0.004).

Conclusions: Single DSH has no effect on IL2SR levels. Schizophrenia has an effect. Multiple DSH has a greater effect. We have insufficient numbers to fully establish the extent of the interaction between schizophrenia and DSH on IL2SR levels. Deliberate self-harm may be a previously overloooked confounding variable in studies on sIL2R in schizophrenia.


< span>Emilia L. Oleszak, Ewa Zaczynska, Meena Bhattacharjee, Catalin Butonoi, Agustin Legido & Christos Katsetos. Fels Inst. for Cancer Res & Mol Biol, Depts Biochem & Neurol & Dept. Microbiol & Immunol, Temple Univ School of Medicine, Philadelphia, PA; Dept pathol, Baylor College Medicine, Houston, TX; Rocky Mountain MS Ctr, Englewood, CO; Sect Pediatric Neurol, St. Christopher’s Hosp Children, Allegheny Univ Health Sciences, Philadelphia, PA

< span>We have examined the localization of inducible-nitric oxide synthase (iNOS) and nitrotyrosine (the product of nitration of tyrosine by peroxynitrite, a highly reactive derivative of nitric oxide (NO) in demyelinating lesions from (i) two young adult patients with acute multiple sclerosis (MS), (ii) a child with MS consistent with diffuse sclerosis, also known as Schilder disease, and (iii) five adult patients with chronic MS. Previous reports have suggested a possible correlation between iNOS, peroxynitrite, related nitrogen-derived oxidants and the demyelinating processes in MS. We have demonstrated iNOS-immunoreactive cells in both acute MS, iNOS was localized in both monocytes/macrophages and reactive astrocytes. iNOS labeled cells of monocyte/macrophage lineage were demonstrated primarily in perivascular lesions. However, foamy (myelin-laden) macrophages, and the majority of reactive astrocytes were iNOS-negative. In the childhood MS case, iNOS protein was present only in a subpopulation of reactive/hypertropic astrocytes. In contrast, no iNOS staining was detected in chronic MS. Immunohistochemical staining of acute MS lesions with an antibody to nitrotyrosine revealed codistribution of iNOS-positive and nitrotyrosine-positive cells, although nitrotyrosine staining was more widespread in cells of the monocyte/macrophage lineage. In diffuse sclerosis-type lesions, nitrotyrosine staining was present in hypertrophic astrocytes, whereas it was absent in chronic MS. In diffuse sclerosis-type lesions perivascular predilection for nitrotyrosine-positive astrocytes was observed. These results suggests that NO and nitrogen-derived oxidants may play a role in the initiation of demyelination in acute MS lesions, but not in the later phase of the disease.


< span>J.A. Engel, J. Zhang, T. Bergström, and L. Svensson. Departments of Pharmacology and Clinical Virology, Göteborg University, Göteborg, Sweden.

< span>One of the current hypothesis regarding the aetiology and pathophysiology of schizophrenia is that it is a neurodevelopmental disorder in which early achieved brain dysfunctions interact with environmental events including viral infections causing schizophrenia in the adult. We have used the prepulse inhibition of acoustic startle experimental paradigm to investigate the long term consequences of neonatal viral infections on the development of this measure of sensorimotor gating function in rats. A marked reduction in prepulse inhibition has been shown in schizophrenic persons suggesting a reduced sensorimotor gating function in these persons. In the present series of experiments, the effect of postnatal inoculation of herpes simplex virus type 1 (HSV-1) was investigated on the development of prepulse inhibition in the rat. This virus was chosen for its neuroinvasive properties and recent observations suggesting the involvement of this virus in the pathophysiology of schizophrenia. HSV-1 inoculated at day 2 after birth, but not at day 4, resulted in a significant delay in the development of prepulse inhibition. A reduction in prepulse inhibition was shown at 37, 46, and 58 days of age in these rats. No evidence was obtained for other behavioral dysfunctions such as differences in sensorimotor reactivity, sensorimotor response habituation, spontaneous locomotor activity, rearing activity or stereotyped behaviour. After completion of the experiments HSV-1 DNA was detected in the substantia nigra of 1 of 3 rats inoculated with HSV-1 on postnatal day 2 indicating persistent/latent infection in these animals. All other samples were negative. The present results suggest that early viral infection may interfere with the development of normal sensorimotor gating function in the rat.


< span>B. Owe-Larsson, T. Andersson, K. Kristensson and R.H. Hill. Stanley Foundation European Research Center, Dept. Neuroscience, Karolinska Institutet, Stockholm, Sweden

< span>Slow, progressive infections of the brain have been reported following mumps virus infections and nervous system dysfunctions as sequelae to mumps encephalitis are not rare. The factors that trigger the onset of the progressive diseases and the mechanisms that lead to neuronal dysfunctions in these as well as in other persistent virus infections remains to be clarified. In this study, the effects of a mumps virus infection on functional properties of embryonic hippocampal neurons in culture were analysed with special emphasis on voltage-dependent Ca2+ channels. Cultures with higher or lower density of glial cells (not treated or treated with mitotic inhibitor, respectively) were infected with the relatively non-cytolytic RW strain of mumps virus and currents were recorded from neurons using whole cell voltage clamp. More than 65% of neurons and glial cells contained viral antigens 1-2 days post infection (p.i.). Glial cells remained infected 6-7 days p.i., while the ratio of infected versus uninfected neurons, especially in cultures with higher glial cell density, was reduced. In both infected and uninfected cultures the somal voltage-dependent Ca2+ currents were stronger in cultures with a higher glial cell density, which indicates that these currents are influenced by glial cells. Introduction of the virus into cultures caused a selective decrease in inward Ca2+ currents, which was most marked at days 6-7 p.i., and which included both infected and uninfected neurons. Spontaneous synaptic currents and other ion channel conductances appeared normal in the infected cultures. Taken together the results show that a mumps viral infection can selectively alter the number or function of some voltage dependent Ca2+ channels in immature hippocampal neurons and that this may reflected a disturbed glial – nerve cell interaction. Such disturbances in the development of a differentiated neuronal function may enhance the susceptibility of the neurons to a second insult later during life even after disappearance of the virus. Supported by a grant from the Stanley Research Award Program.


< span>A.G. Alias. Chester Mental Health Center, Chester, IL 62233

< span>Defective information processing, poor working memory, inability to quickly shift attention without fumbling, ambivalence, avolition, anhedonia, and impaired motor coordination are cardinal features of schizophrenia but, unlike delusions and hallucinations, they are related more to negative/deficit symptoms. As summarized by Bass, numerous studies have correlated leadership with “speed and accuracy of thought,” “finality of decision” (the opposite of ambivalence), “ambition, initiative and persistence ” (opposite of avolition), “mood control, and sense of humor” (opposite of anhedonia) etc. Torre wrote “political leaders have certain personality. . .traits [that] enable them to seek out and be cast in a leadership role. The first….is high energy… They are able to process large quantities of information quickly…[Curiously, they] also have enormous sexual appetites.” Physical hardiness, as in Ronald Reagan or Fidel Castro, is another trait too often noted in leaders. Goodwin and Jamison have demonstrated a relationship between leadership and hypomania.

< span>High sex drive in leaders could be a reflection of hypomania, but when coupled with physical hardiness also, it must be a reflection of higher androgen effects as well. Further, Gray et al have found positive correlations between dominance and serum levels of dehydroepiandrosterone (DHEA) and testosterone (T), but not of more potent dihydrotestosterone (DHT), in over 1,700 older men. While DHEA potentiates certain immune system profiles, DHT suppresses some of them; T’s role is more complex, maybe because T is an immediate precursor of DHT and estradiol.

< span>A relationship between cognitive processes and cerebral and basal ganglia functions, and a role of neocerebellum in rapidly shifting attention, have been demonstrated. The cognitive styles, including a proficiency to quickly shift attention, of John Kennedy, Napoleon, and Julius Caesar are used as examples of the contrasting model. Caesar was known to dictate to four to seven secretaries simultaneously. Mustapha Kemal could “sip coffee, smoke, hold a liquor glass, and look at pictures – all at the same time.”

< span>It is suggested that specific brain imaging studies could illustrate this contrast. Therapeutically, prior to 1950s, DHEA had been used in young schizophrenics with modest success in improving deficit symptoms. DHEA and its derivatives may prove to be valuable to treat deficit symptoms of schizophrenia in both sexes.


< span>John L. Black, Elaine M. Barth, Don E. Williams, Joyce A. Tinsley. Mayo Foundation, Mayo Clinic, Rochester, MN

< span>Stiff-man syndrome (SMS) is a rare autoimmune neurologic disorder associated with antiglutamic acid decarboxylase (anti-Gad-65) autoantibodies. Thirteen patients with SMS were studied with the Minnesota Multiphasic Personality Inventory (MMPI), the Self-Administered Alcoholism Screening Test (SAAST), the State-Trait Anxiety Inventory profiles (STAI), and by telephone interviews. The mean MMPI, SAAST and STAI were within normal limits; several patients had abnormal profiles. The results of telephone interviews revealed that eight patients (62%) had been given at lease one psychiatric diagnosis and four (31%) abused alcohol or were dependent upon it. Two patients had a psychiatric diagnosis that preceded the onset of symptoms of SMS. We hypothesize that SMS patients have a gamma amino-butyric acid deficiency or GAGAergic neuron dysfunction as a result of the anti-GAD-65 autoantibodies that leads to psychiatric symptoms including depression and chemical abuse. This naturalistic study raises interesting questions as to the etiology of depression and chemical dependency in some patient populations.

< span>

Plenary Speaker


< span>John F. Kurtzke. Neuroepidemiology Section, Neurology Service, VAMC, Washington DC

< span>From over 300 community based surveys about the world, prevalence rates for multiple sclerosis can be grouped into three zones. High frequency areas with prevalence of 30 or more per 100,000 population comprise Europe, Canada, northern United States, New Zealand, southeastern Australia, and easternmost Asian Russia. These are bounded by medium frequency areas with prevalence of 5 to 29. Low rates (under 5) are found in much of Asia and Africa, the Caribbean region and northern South America. Regardless of geography MS is most common among whites.

< span>Migration between different frequency areas changes the risk of MS. High to low migrants retain high risk only if they move at age 15 or older. Residents moving into high risk areas between 11 and 45 or so increase their risk.

< span>Epidemics of MS during the World War II period have been described for the Faroe Islands, Iceland, and the Shetland-Orkney Islands. Best insight as to the nature of MS comes from the Faroes where the disease was unknown until 1943. This first epidemic with maximal incidence in 1945 was followed by three later epidemics, each one peaking 13 years after the previous.

< span>The Faroes were occupied by British troops 1940-1945. We believe they brought a specific, persistent, widespread infection which was transmitted to the Faroese population cohort then age 11 to 45, and part of this infected cohort later transmitted it to the next cohort of Faroes as it reached the susceptible age of 11. This infection we call the primary multiple sclerosis affection (PMSA) to distinguish it from its rare late sequel of clinical neurologic MS (CNMS). Prolonged exposure is needed to acquire PMSA, which itself could well be a novel (retro) virus.

< span>


< span>AS Brown, ES Susser, P. Cohen, WH Hoek. New York State Psychiatric Institute.

< span>We shall describe preliminary results from two recent investigations of affective disorders following prenatal viral exposures. The first is a study of a birth cohort prospectively documented with in utero rubella, previously found to have an increased risk of schizophrenia in adulthood. This cohort consisted of 70 individuals in whom clinical and serological confirmation of prenatal rubella infection was obtained, and who received a standard psychiatric diagnostic interview during a follow-up assessment in young adulthood. In the present study, we examined whether prenatal rubella exposure also resulted in an increased risk of affective disorders (unipolar major depression and bipolar affective disorder). We found that the risk of affective disorder was similar between the rubella-exposed cohort and the unexposed cohort [RR (95% CI)=1.1 (0.65-1.9), p=0.661]. This finding supports the specificity of the association between prenatal rubella and schizophrenia among major psychiatric disorders. In the second investigation, we examined the risk of hospitalized affective psychosis (ICD-9 296), using psychiatric registry diagnoses, following exposure to the 1957 A2 influenza epidemic in Holland. In males, the relative risk (RR) of affective psychosis following influenza exposure in the sixth month of gestation was increased nearly twofold in makes [RR (95% CI)=1.78 (1.12-2.82), p=.01], but not in females [RR (95% CI)=.74 (.49-1.11), p=.15]. These data suggest that prenatal influenza exposure may play a role in the etiology of affective psychosis in males. Taken together, the findings from these two studies indicate that the effect of prenatal viral exposure on risk of affective disorder may be modified by the type of viral agent, the timing of infection, and gender.

< span>


< span>M. Clarke, F. Keogh, M. Morris, C. Larkin, D. Walsh, E. O’Callaghan. Stanley Research Unit, Cluain Mhuire Family Centre, St. John of God Adult Psychiatric Services, Newtownpark Avenue, Blackrock Co Dublin, Ireland; the Health Research Board, 73 Lower Baggott Street, Dublin 2, Ireland

< span>Background: Although seasonality of admission has been studied extensively in affective disorder, less work has been done in schizophrenia. We examined variation in the admission patterns of schizophrenia and affective disorder both cumulatively, and year by year, using several statistical methods.

< span>Methods: Using the national Psychiatric Inpatient Reporting System, all individuals admitted between the years 1989-94 with an ICD 9/10 diagnosis of a first episode of schizophrenia or affective disorder were identified. Seasonal variation in their admission patterns was examined and using the chi square test and a Kolgomorov-Smirnov type statistic, and then by applying the Walter and Elwood method. Interactions between variables were assessed by using a hierarchical log linear model.

< span>Results: Significant seasonal variation existed in the monthly admission patterns of both schizophrenia (c2 = 18.28, DF = 2, P<.001) and affective disorder (c2 = 39.43, DF = 2, P<.001), with summer peaks. The effect was most marked for the subgroup with mania (c2= 49.43, DF = 2, P<.001). A significant interaction was found between season of admission and admission year in schizophrenia (change in scaled deviance 11.1437, DF = 5, P<.0485). The magnitude of the July excess was 12% in schizophrenia and 17% in mania.

Conclusions: Although individuals with mania showed the most pronounced summer peaks, first admissions of schizophrenia are influenced by seasonal factors. Potential candidates include infections, climatic factors or toxins.

This study was funded by the Stanley Foundation.


< span>TA Klempan, P Deb, RL O’Reilly, EF Torrey, and SM Singh. Molecular Genetics Unit, Department of Zoology, Psychiatry, and Division of Medical Genetics, The University of Western Ontario, London, Ontario, Canada; NIMH Neurosciences Center at St. Elizabeth’s, Washinton DC

< span>Representational difference analysis (RDA) is a molecular technique which facilitates the identification of differences between two paired samples of DNA. We have employed RDA towards the analysis of differences in retroviral-like sequences between monozygotic twins discordant for schizophrenia. Comparing schizophrenic patients and unrelated controls we observed a distinct increase in amplification of retroviral reverse transcriptase (Pol gene) amplicons over five successive rounds of RDA. In contrast, there were no clear differences of this type observed with a pair of discordant monozygotic twins over a similar number of rounds of RDA. All RDA products sequenced from these experiments were confirmed to be of retroviral origin, thus demonstrating the integrity of the technique in only amplifying DNA of retroviral origin. Of interest, we note in the discordant twin pair one sequence which was seen only at the fifth round in the different combination and possesses homology to a known ERV FTD putative reverse transcriptase (pol) pseudogene clone (81%), the human retrovirus-like sequence-isoleucine a (RTVL-IA) gene (77%) and several human haptoglobin protein genes. Another sequence has strong resemblance (86% to 95% matches) to a number of recently characterized multiple sclerosis associated retrovirus (MSRV) pol gene sequences. Transcripts of these sequences, related to the endogenous element ERV9, have recently been detected from serum and tissue culture of multiple sclerosis patients while relatively absent from control subjects. A further product from the first round of the different combination shows an extremely high degree of homology to the 5′ end of a known human poly (ADP-ribose) polymerase gene (96%). The identity of this sequence is virtually certain and may indicate a site of ancestral retroviral integration. Other more common sequences have been noted with identity to a cloned human ERV9 reverse transcriptase homolog isolated from brain (91%), several other human endogenous retrovirus pHE.1 mRNA sequences (73%) and other MSRV pol gene sequences. Continued analysis of this twin pair and other discordant twins is necessary to confirm these findings. If confirmed we will undertake stringent screening of these sequences to determine possible association with schizophrenia.


< span>W. John Martin and Donovan Anderson. Center for Complex Infectious Diseases, Rosemead, CA 91770

< span>A major ongoing epidemic of neuropsychiatric illnesses in the Mohave Valley region of the United States has been linked to a stealth-adapted virus. The infection spreads easily between family members and has occurred in household pets. A physician caring for these patients also became infected. Although, molecularly heterogeneous, stealth viruses produce a generally similar cytopathic effect (CPE) in both human and animal cell lines. The CPE in MRC-5 cells begins with a rounding of the normal elongated spindle shaped fibroblasts. The enlarged cells adhere to form tight clusters, often with syncytia formation. The cytoplasm acquires a foamy, vacuolated and slightly granular appearance. These changes are especially prominent in tests performed on patients from the Mohave Valley (Pathobiology 65:51-56, 1997).

< span>Blood and cerebrospinal fluid (CSF) from Mohave Valley patients are readily distinguished from controls in repeated double-blind viral culture studies (e.g. in one such study 18/19 Mohave Valley patients were strongly positive versus 1 of 8 patients from a Los Angeles community hospital who gave a weak positive CPE).

< span>The Mohave outbreak has helped define the clinical spectrum of illness associated with stealth viral infections and has provided useful insights into various dysfunctional brain syndromes. In particular, the striking neurological manifestations of a subset of severely ill patients and the positive culture results on the generally acellular CSF, are consistent with a non-inflammatory brain infection disrupting normal brain function. Several stealth virus culture positive patients have presented with a primary psychiatric illness, including manic depression and psychosis. The majority of patients have symptoms generally associated with the chronic fatigue syndrome (CFS). Non-neurological manifestations seen in some of the patients include gastrointestinal, hepatic, endocrine and cardiac diseases. The clinical terms stealth virus infection with, or without, encephalopathy (abbreviated SVIE and SVIE, respectively) have been proposed.

< span>


< span>D.J. Hart, R.G. Heath, C.H. Methvin, J. Castillo, and T.J. Briggs. Departments of Psychiatry and Neurology; The School of Medicine; MEdREP, Tulane University, N.O., LA

< span>One of the biological hypotheses concerning the etiologies of schizophrenia and bipolar disorder proposes that the conditions are the consequence of a virally-induced autoimmune disease. A detailed analysis of the results obtained from screening archived serum samples for retroviral antibodies revealed significant reactivity among some patients having either schizophrenia, a schizophrenia-spectrum disorder, or bipolar disorder. In further investigation of this hypothesis, Western blots were prepared from the septal region of a Rhesus monkey brain. Several samples from schizophrenics who were reactive on HIV-1 Western blots, one schizophrenic blot-negative sample, and five blot-negative samples from the blood bank were tested for reactivity against the septal antigens. The preliminary results suggest a positive correlation between the presence of certain anti-retroviral and anti-septal antibody reactivity in the sera of the schizophrenics tested. An immunologic cross-reaction could account for these results and would be consistent with the viral/autoimmune hypothesis.


< span>Janice R. Stevens. Oregon Health Sciences University, Portland, Oregon 97201

< span>Epilepsy and schizophrenia are two of the most common disorders of the nervous system. Thanks in part to modern pharmacology, 60-80% of patients diagnosed with these disorders can not be rendered symptom free with proper treatment. The drugs used for their treatment may provide clues to their etiology or pathophysiology. Drugs used to treat epilepsy are primarily directed toward reducing hyperexcitability in brain neurones. the most potent of these agents, the competitive and non-competitive excitatory amino acid antagonists cause psychoses and for this reason are not used clinically. Drugs used to treat schizophrenia, on the other hand, generally oppose one or more inhibitory networks in the brain, and all can cause epileptic seizures. Clozapine, one of the most effective of the modern antischizophrenic drugs, is the most epileptogenic of them all, causing seizures in 3-5% of those treated with this agent.

< span>The brain is composed of complex networks of excitatory and inhibitory transmitters, pre and post-synaptic receptors and uptake sites that regulate the equilibrium between excitation and inhibition at output sites. Specific neurotransmitters, amino acids, monoamines, peptides and hormones interact with a large variety of very specific receptors, which together determine whether the output of a given network will be excitatory or inhibitory. Loss of the physiologic balance between these factors in specific networks yield overactivity and epilepsy on the one hand or excessive inhibition and psychosis on the other. Pooling and averaging data from large numbers of individuals diagnosed with schizophrenia who have diverse symptoms, course and response to pharmacologic agents, may in schizophrenia, as it would in epilepsy, obscure factors which ar closer to the etiology of the disorder. Alternative strategies are proposed for analyses of the diverse data from schizophrenia studies.

< span>


< span>Maree J. Webster, Robert H. Yolken, E. Fuller Torrey, Linda D. Bobo. Stanley Foundation Research Program, NIMH Neuroscience Center, Washington DC; Stanley Neurovirology Laboratory, Johns Hopkins University, Baltimore, MD

< span>Epidemiologic studies suggest that infectious or immune events may play a role in the etiology of serious neuropsychiatric diseases, such as schizophrenia. Cytokines, such as interleukin-2, which may be induced by infection, are pluripotent with regard to immune and CNS function and behavior. Although IL-2 binding of its trimolecular cell surface a b g receptors initiates these complex series of events, only the a receptor is inducible in response to continued antigenic stimulus. Bobo, et al, recently described significantly elevated levels of interleukin-2 receptor a (IL-2Ra) in the CSF of patients with schizophrenia and bipolar disorder with psychosis as compared to nonpsychotic unipolar depressed and normal controls. To determine if IL-2Ra is significantly elevated in particular brain areas in association with elevated CSF levels immunoperoxidase staining was performed on post-mortem sections of frontal cortex, striatum, and hippocampus of patients with high (>200 pg/ml), medium (100-200 pg/ml) and low (<100 pg/ml) CSF IL-2R as determined by ELISA. In a preliminary study, we found no evidence of IL-2Ra protein in the frontal cortex of any subject, thus supporting the results of Bobo, et. al that showed neither of IL-2R protein by ELISA nor IL-2R mRNA by RT-PCR in the frontal cortex. Similarly, IL-2R was not found in the striatal sections. However, cases exhibiting high levels of IL-2R in the CSF showed positive immunostaining for Il-2R in the subependymal cells of the lateral ventricle. One schizophrenic case also exhibited some staining in the choroid plexus and of the pyramidal cells of the hippocampal formation. In contrast, there was no evidence by immunohistochemistry or in situ hybridization of tissue IL-2 cytokine. In order to support the hypothesis that there could be continued immune activation in association with expression of Il-2Ra, we are also performing co-localization studies for MHC Class II complex and for cytomegalovirus immediate/early proteins, since CMV target subependymal cells.


< span>E.L. Oleszak, W. L. Lin, J. Kuzmak, J.R. Chang, E. Zaczynska and C.D. Platsoucas. Fels Inst. for Cancer Res & Mol Biol and Depts of Biochem & Neurol, & Dept Microbiol & Immun, Temple University School of Medicine, Philadelphia, PA

< span>Theiler’s murine encephalomyelitis virus (TMEV) causes in susceptible strains of mice biphasic disease in the CNS, consisting of early acute disease that resembles polioencephalitis and of late chronic demyelinating disease. Both are associated with the presence in the CNS of extensive mononuclear cell infiltration primarily consisting of T cells and monocyte/macrophages. TMEV-induced chronic demyelinating disease in mice is an excellent model of multiple sclerosis. The pathogenesis of TMEV-induced demyelinating disease of the CNS is poorly understood. There is no information on the antigenic determinants, host or viral, recognized by T cells infiltrating the CNS of TMEV-infected mice with late chronic demyelinating disease. We amplified using the nonpalindromic adaptor PCR ß-chain T-cell receptor (TCR) transcripts from the spinal cord of SJL mice with late chronic demyelinating disease (170 and 192 days postinfection [p.i.]). We cloned and sequenced the amplified TCR transcripts. Sequence analysis revealed the presence of substantial proportions of identical ß-chain TCR transcripts, suggesting the presence of monoclonal T cells in the spinal cords of TMEV-infected SJL mice with late chronic demyelinating disease. Two clonally expanded TCR transcripts were found in these mice, the Vß2.1 Dß2.1 Jß2.1 (in mice 170 days p.i.) and Vß2.1 Dß2.1 Jß2.4 (in a mouse 192 days p.i.). The two ß-chain TCR sequences are novel, have not been previously identified, and were not in the GENBANK/EMBL database. Confirmation of the clonally expanded Vß2.1 Dß2.1 Jß2.1 transcript was obtained by two-sided PCR, using a 5′ Vß2.1 primer and a 3′ Cß primer, followed by cloning and sequencing. 7% of the ß-chain transcripts found in the spinal cord of TMEV-infected mice with late chronic demyelinating disease, employed CDR3 TCR motifs homologous to those previously found in MBP-specific T cells. Although these transcripts were not clonally expanded, their presence in the CNS of these mice is not physiological, and may be relevant to the pathogenesis of demyelinating disease. The Vß2.1 Dß2.1 Jß2.1 TCR sequence that was clonally expanded in the CNS of SJL mice with late demyelinating disease was also clonally expanded in the CNS of SJL mice with early acute disease, 8 days p.i. These results demonstrate that a particular ß-chain TCR transcript (Vß2.1 Dß2.1 Jß2.1) is clonally expaned in the spinal cord of TMEV-infected SJL mice with late chronic demyelinating disease and in the CNS of mice with early acute (8 days p.i.). T cells utilizing this transcript may play a role in the pathogenesis of the disease.

< span>



< span>Lothar Stitz. Institute for Vaccines, Federal Research Center for Virus Diseases of Animals, Tjbingen, Germany

< span>In experimentally infected, as well as in naturally infected animals, Borna Disease Virus (BDV) infection generally results in an encephalomyelitis associated with severe neurological symptoms in the acute phase. Prior to the onset of encephalitis we have observed disorders of the cortical cholinergic innervation which might be responsible for early behavioral and mood disturbances. In the chronic phase signs of dementia and chronic debility are prevailing symptoms associated with a severe cortical brain degeneration. During recent years we have studied the pathogenesis of BD in rats and have found that the disease is based on an immunopathological response to the noncytopathic BDV. CD4+ T cells, CD8+ T cells, macrophages and B lymphocytes are detectable during the phase of inflammation in the brain, but late after infection no reaction can be found. The T cell-mediated immunopathological reaction is caused by CD4+ T cells which have helpers functions and CD8+ T cells which act as effector cells recognizing MHC class I-expression virus infected target cells in the brain. The major pathway of destruction of target cells by CD8+ T cells has been identified in BDV-infected rats by demonstrating the presence of perforin in the brain. Interestingly, the quality of the inflammatory reaction in naturally infected horses and sheep is comparable to the situation found in experimentally infected rats. The intra vitam diagnosis of BD in naturally infected hosts represents a major problem. Antibodies to various virus specific proteins can be detected in the peripheral blood of infected individuals but there is an upcoming debate in how far the presence of antibodies directed against one or more viral proteins can be used as parameters of an infection. In some sheep and cattle we have detected serum antibodies to virus specific proteins but neither infectious virus nor nucleic acid in the brain. Virus specific nucleic acid was not detected in the peripheral blood of immunocompetent experimentally infected or naturally infected hosts. Finally, we extended our search for antibodies or virus to psychiatric patients. Employing a single method we could detect virus specific antibodies but failed to verify the results by using other methods. Likewise, in a blind multicenter study, we were not able to detect virus specific nucleic acid in blood samples of schizophrenic patients by RT-PCR. From these data we conclude that a single test method for the detection of antibodies is inappropriate and that the screening of blood samples to detect the presence of specific nucleic acid is an unsuitable technique to identify BDV-infected neuropsychiatric patients.

< span>


< span>Yasuhide Iwata, Kazuo Takahashi, Xie Peng, Koji Fukuda, Tsuguhiro Ogawa, Norio Mori, Shin-ichi Niwa, Shiro Shigeta. Department of Microbiology, Neuropsychiatry, Fukushima Medical College, Fukushima, Japan; Departments of Psychiatry and Neurology, Hamamatsu University School of Medicine, Hamamatsu,Japan

< span>It has been reported that BDV infection may be associated with human psychiatric disorder. To confirm this association, BDV p24 RNA was detected by nested RT-PCR in PBMCs of psychiatric patients and blood donors in Fukushima, Japan. The prevalence of BDV p24 RNA in patients with major depression, schizophrenia and in blood donors were 6% (2/36), 4% (3/71) and 2% (2/91), respectively. All positive samples were confirmed to be positive by repeating the test, and the samples showing positive twice were regarded as positive. Three to 5 clones from nested RT-PCR products of 2 patients with major depression, 3 patients with schizophrenia, 1 blood donor and BDV/MDCK were sequenced and the human BDV p24 sequences were compared with horse BDV p24 sequences. There were 2 positions of mutation which were common to all clones from one or some of the individuals. One mutation from C to T at nt 1535 with no amino acid change and another mutation from G to A at nt 1617 with a conservative amino acid change (Val®Thr) were found. The divergency of human sequence compared with BDV/MDCK (0-1.79%) was lower than that compared with C6BV (3.58-6.07%) or strain V (1.79-3.28%). Experimentally induced mutations in our nested RT-PCR examined with the RNA from BDV/MDCK as a template showed 3 positions of mutation with cluster at 1669 1705, and 1738 and several point mutations, suggesting the mutation during RT step and the point mutation rate of 0.6% at most. In comparison with human BDV p24 RNA sequences previously reported from Japan and Germany, there were several positions of silent nucleotide mutations which could distinguish geographically different clones, suggesting that there may exist geographical area-specific BDV strains in these countries.


< span>K. Ikeda, S. Haga, M. Yoshimura, K. Arima, M. Kazamatsuri, K. Takahashi, K. Ikuta, M. Tashiro. Department of Ultrastructure and Histochemistry, Tokyo Institute of Psychiatry, Tokyo 156 Japan; Tokyo Metropolitan Matsuzawa Hospital, Tokyo; National Institute of Neurology and Psychiatry, Tokyo; Hokkaido University, Sapproro; National Institute of Infectious Diseases, Tokyo

< span>Borna disease virus (BDV) is a neurotropic RNA virus, naturally infecting horses and sheep in central Europe. Since the seroepidemiological studies in 1985, BDV has been suspected to be associated with human psychiatric disorders, including schizophrenia and mood disorders. Very recently, BDV mRNA for p24 has been detected in autopsy brain samples exclusively from schizophrenia and bipolar disorder, but not in samples from patients with several neurological disorders and control subjects (Salvadore et al, Lancet 349;1813-1814, 1997).

< span>We have also looked for the BDV footprint in autopsy brain samples by using the same highly sensitive nested reverse transcriptase (RT)-PCR. Four brain regions (frontal and temporal cortices, hippocampus, olfactory bulb) were obtained at autopsy. Total RNA extracted from these samples were used for the nested RT-PCR analysis of BDV p24 and p40 genomes. The PCR protocols for p24 genome with the expected size of 374 bp were found in three of 43 controls, and 5 of 23 schizophrenic patients. Some PCR products were verified by direct sequencing and were found to be analogous to the similar portion of the strain He/80-2 of BDV. The PCR products for p40 genome were found in one of 43 controls, and three of 23 schizophrenics. Noteworthy was that not all four regions harbored viral fragments in a single “positive” case. The coexistence of p24 and p40 genome was found in one control individual and one patient.

< span>Our data suggest that BDV can latently infect normal subjects, presumably in brain tissue, and the prevalence of BDV genomes in human autopsy brain samples in Japan might be higher in schizophrenic patients than in controls.

< span>


< span>Masahiko Kishi, Takeshi Kobayashi, Yuko Shoya, Toshiaki Koda, Patrick K. Lai, Kazuyoshi Ikuta, and Mitsuaki Kakinuma. Institute of Immunological Science, Hokkaido University, Sapporo 060, Japan; Department of Biosciences, Salem-Teikyo University, Salem, West Virginia 26426

< span>The Borna disease virus (BDV) replicates in the nucleus. The viral nucleoprotein (N), which is found abundantly in the nucleus in BDV-infected cells, may play an important role in the virus replication. To examine the amino acid region involved in the nuclear targeting of BDV N, a series of eukaryotic expression plasmids encoding various deletion mutants of N was constructed and transfected into COS-7 cells. In indirect immunofluorescence assay with a rabbit anti-BDV N antiserum, a wild-type was localized in the nucelus of cells transfected in the absence of other viral constituents. In contrast, mutants commonly lacking the 13 N-terminal amino acid region (M1PPKRRLVDDADA13) gave a cytoplasmic localization pattern. Similarly, a mutant with substitution of K4RR6 to N4TG6 was retained in the cytoplasm. To further analyze the nuclear localization activity, the octapeptide (P3KRRLVDDA11) was fused to ß-galactosidase. After transfection, the fusion protein was found in the nucleus, whereas the ß-galactosidase control was found in the cytoplasm. Thus, the 9 N-terminal amino acid region is required for the nuclear targeting of BDV N and the K4RR6 motif is essential for its nuclear targeting activity.


< span>Tahir H. Malik, Takeshi Kobayashi, Yuko Shoya, Masahiko Kishi, & Patrick K. Lai. Department of Biosciences, Salem-Teikyo University, Salem, WV 26426 USA; Inst. Immunological Science, Hokkaido University, Sapporo 064, Japan

< span>Previous studies predicted the presence of a small open reading frame (ORFx1) located between ORF-1 and ORF-2 of the Borna Disease viral (BDV) genome. We have cloned the nucleotide sequence of ORFx1 into expression vectors. Transfection of Cos-7 cells with ORFx1 tagged to FLAG gave cytoplasmic staining with an anti-ORFx1 serum. This anti-serum did not react with the Cos-7 or the MDCK cells, but gave nuclear and cytoplasmic fluorescence when used to stain BDV-infected MDCK cells. We hypothesized that the ORFx1 protein (p10) might interact with other BDV protein(s) which carries a nuclear localization signal. We performed in vitro assays to determine whether the p10 can bind to the BDV p40 nucleoprotein, or the p24 phosphoprotein. We cross-linked GST-p40 or GST-p24 to solid phase and studied the interaction with in vitro transcribed and translated p10-FLAG. Interaction of p10 and p40 and p24 may be a mechanism whereby the protein becomes localized in the nucleus where BDV replicates. Likely, p10 is a regulatory protein analogous to the C-protein of the Sendai virus.

< span>


Carolyn G. Hatalski, William F. Hickey and W. Ian Lipkin. Laboratory for Neurovirology, University of California-Irvine and Department of Pathology, Dartmouth Medical School

Immunocompetent Lewis rats infected with Borna disease virus exhibit a multiphasic immune-mediated disease of the central nervous system. Mononuclear cell infiltration is marked within the first few weeks of infection but later subsides despite the continued presence of infectious virus, viral proteins and the viral transcripts. In an effort to understand the basis for changes in the immune response and implications for pathogenesis and viral persistence, we examined immune cell infiltrate and cytokine mRNA profiles in infected rats throughout the course of infection. In the acute phase, perivascular immune cell infiltrates consisted of CD4+ and CD8+ T cells, macrophages and NK cells. In the chronic phase, the numbers of perivascular T cells and macrophages declined although the number of NK cells, B cells and activated microglia increased in the brain parenchyma. The reduction of immune cell infiltrates was associated with the presence of terminal deoxynucleotidyl transferase dUTP-biotin nick end labeled (TUNEL) cells suggesting apoptosis of infiltrating inflammatory cells. The profile of cytokine mRNAs showed a peak expression of interleukin-1, alpha, interleukin-2, interleukin-6, tumor necrosis factor alpha and interferon gamma in the acute phase of disease, coincident with the peak of immune cell infiltration. In contrast, interleukin-4 mRNA, detected in cells with morphology consistent with lymphocytes, was expressed at highest levels in the chronic phase of disease. Expression of transforming growth factor beta mRNA was elevated throughout infection. Analysis of sera from rats at different times post infection revealed increased IgA in the acute phase and increased IgG2a, IgG2b and IgE in the chronic phase of infection. Elevated levels of IL4 mRNA and IgE serum antibodies in the chronic phase of infection are consistent with a switch from a Th1-like, cellular immune response in the acute phase of infection to at TH2-like, humoral immune response in the chronic phase of infection. This shift may provide a mechanism for limiting destruction of neural tissue by inflammatory cells.


< span>Carolyn G. Hatalski, Martin Schwemmle, Mady Hornig, W. Ian Lipkin. Laboratory for Neurovirology, University of California-Irvine and Mood and Anxiety Disorders Section, Department of Psychiatry, University of Pennsylvania

< span>Borna disease (BD) virus is a noncytolytic, neurotropic virus that causes a persistent infection of the nervous system. Infection of the adult, immunocompetent Lewis rat first causes a cellular immune response in the central nervous system (CNS) that results in a dramatic neurologic syndrome. In the chronic phase of the disease cellular infiltration recedes; however, loss of brain mass continues. Subtractive cloning experiments were pursued to explore the differences in host gene expression in the CNS during the acute and chronic phases of BD. Representational difference analysis (RDA) revealed enhanced expression of immunoglobulin and ferritin mRNAs in the chronic BD rat brain. In situ hybridization experiments revealed a marked infiltration of plasma cells. Analysis of serum and cerebral spinal fluid (CSF) confirmed that BDV-specific antibodies, including antibodies that neutralize viral infectivity, were concentrated within the CSF. Comparisons of the IgG isotypes from the brain parenchyma of rats at different stages of infection revealed peak levels of IgG2a in the acute phase with elevated levels of IgG1, IgG2b and IgG2c in the chronic phase. These data support our finding of a Th1 to Th2 shift in BD and suggest that a humoral mechanism may be implicated in the pathogenesis of CNS damage during the latter phase of BD.

< span>


< span>K. Fukuda, K. Takahashi, Y. Iwata, T. Horimoto, M. Tashiro, N. Mori, S. Niwa and S. Shigeta. Department of Microbiology and Neuropsychiatry, Fukushima Medical College, National Institute of Infectious Diseases, Department of Psychiatry and Neurology, Hamamatsu University School of Medicine

< span>It has been suggested that some of the patients with psychiatric diseases may be infected with Borna disease virus (BDV). In the present study we have tested the T cell responses to BDV proteins in psychiatric patients to examine whether the cellular immune response to BDV could be detected and whether T cell responses could be one of the appropriate diagnostic test as well as the specific antibody and BDV RNA. The specific T cell response was examined by proliferation assay of PBMC stimulated with recombinant p24 or p40 BDV protein which was purified by Glutathion-sepharose and Mono-Q column chromatography. The antibody to p40 was tested by Western blot assay and BDV p24 RNA in PBMC was examined in nested RT-PCR. Positive proliferative responses to both p24 and p40 were detected in 10% (3/30) of schizophrenic patients, 13.3% (4/30) of mood disorders and 6.7% (2/30) of blood donors. There was no significant difference in the prevalence between each group. Antibody to BDV p40 was detected in 3.6% (1/28) of schizophrenic patients, 11.4% (4/35) of mood disorders and 0% (0/32) of blood donors. BDV RNA was detected in 7.7% (2/26) of schizophrenic patients, and 15% (2/26) of mood disorders and 0% (0/31) of blood donors. The positive proliferative responses in these populations suggests that it may be due to BDV infection or response to cross-reactive unknown antigens. Two out of 4 mood disorder patients who were positive for the proliferative response showed the very high proliferative response with stimulation index of over 10. One of the two patients was also positive for both antibody to BDV p40 and BDV p24 RNA, indicating that this patient is definitely infected with BDV. Evaluation of the results of combination of these three assays will be discussed.

< span>


< span>KM Carbone, SA Rubin, RW Waltrip. Laboratory of Pediatric and Respiratory Viral Diseases, FDA

< span>There is an urgent need for a sensitive and specific diagnostic assay for BDV infection of humans. For years, variations of the indirect immunofluorescence assay (IFA), utilizing BDV-infected MDCK or C6 cells, predominated for diagnosis of BDV infection. The IFA is quick and sensitive, and remains useful as an experimental tool (i.e., when the populations of infected and uninfected animals are known). However, the subjective, low-specificity IFA has limited use for screening of human populations with a low prevalence of seropositives. The time consuming Western blot assay (WB), used to screen horse, rat, mouse and human sera, is sensitive and more specific that the IFA, IF stringent criteria were utilized (defining seropositive as the recognition of TWO or more different BDV proteins). Rapid ELISA-type BDV serological assays are in use, using bacterial or insect cell recombinant antigens with variable success. While the ELISA can detect anti-BDV antibodies in rats, when utilized for experimentally-infected rabbits and horses, and humans (unknown), the ELISAs have not been as convincingly useful. Still, when compared to RT-PCR, serological diagnosis of BDV infection may be the preferred method, given evidence of seropositive subjects with negative PBMC RT-PCR results. Moreover, there are other limitations of the use of RT-PCR as a screening tool given the small numbers of BDV-infected PBMCs and evidence of false positive results in some laboratories.

< span>There are likely to be a multitude of reasons why BDV serology results are variable and inconsistent. In the cases of horses, rabbits, and perhaps humans, when tests of high specificity are used the antibodies appear to be of low avidity and in low titers. In addition, conformational epitopes and unstable protein antigens have complicated developmental of a serological test. Here we review the serological data from several experimental systems and present new data which illustrate the difficulties in designing a sensitive and specific anti-BDV serological screening assay for use in humans.

< span>



< span>Paula Rantakallio, Peter Jones, Juha Moring, and Lennart von Wendt. University of Oulu, Finland; University of Nottingham, U.K., Bracke Ostergard Regional Pediatric Rehabilitation Centre, Gothenburg, Sweden

< span>Background: Maternal exposure to influenza epidemics during pregnancy may increase the risk of schizophrenia in the off-spring. We investigated the association between central nervous system (CNS) infections defined prospectively up to the age of 14, and later onset of schizophrenia and other psychoses in the 1966 birth cohort in Northern Finland, which covers 96% of all births in the area during that year.

< span>Method: Data regarding CNS infections were collected in 1966-1980. Registered diagnoses of psychoses in 1982-1993 were validated in terms of DSM-III-R criteria.

< span>Results: Of 11,017 subjects alive at 16 years, 145 had suffered a CNS infection during childhood, 102 of them a viral infection. 76 had DSM-III-R schizophrenia and 53 some other psychosis. 4 cases of schizophrenia had suffered viral CNS infection and 2 cases of other psychosis a bacterial infection. When major neurological abnormalities and fathers social class were adjusted, the odds ratio (OR) of schizophrenia after viral CNS infection was 4.8 (CI 1.6, 14.0) the other significant risk factors being intelligence quotient (IQ) <85, perinatal brain damage and male sex, but not epilepsy. Similarly, adjusted OR of other psychoses was 6.9 (CI 1.4, 32.8) after bacterial CNS infection, the other significant risk factors being IQ <85 and severe hearing defect. Two of the four viral infections were caused by Coxsackie B5 during an epidemic in which 16 neonates were infected together.

< span>Conclusions: CNS infections during childhood clearly carried an increased risk of adult onset schizophrenia or other psychoses, viral infections being important for schizophrenia, particularly Coxsackie B5 during the new born period.