A Positional Cloning Study on Schizophrenia

• Specific Aims This component project of positional cloning study of schizophrenia (POCOS) has two specific aims: (1) To locate and identify the vulnerability genes of the phenomenological phenotype as well as the endophenotypes of schizophrenia (SCH) at specific chromosome regions; (2) To find specific polymorphism of candidate genes associated with endophenotypes and/or phenomenological phenotypes of schizophrenia. These results will lead this POCOS team: (1) to develop new clinical diagnostic method and new prevention program including pharmacological agent in early intervention treatment of SCH for public health purpose; (2) to do functional genomic study on SCH; (3) to study the pathogenetic process of abnormal genes in abnormal neuropsychological and neurobiological functions of SCH; (4) to delineate the nature and the effect of gene-environmental interaction in the etiology of SCH.

Up to the present time, all genome-wide scans for localization of vulnerability genes revealed no consistent results. The difficulty of molecular genetic study on schizophrenia is not the technology of molecular genetic study; It is the difficulty in clinical recruitment of adequate samples. Small sample size, diagnostic uncertainty, and multiple ethnicities of study samples were major reasons for this present unfruitful condition. This POCOS is designed with a remarkable strength of sample characteristics: (1) Using enough big sample size with DNA sample of around 700 families with at least two schizophrenic siblings. This family sample assures adequate power for linkage analysis and further positional cloning strategy, (2) Using standardized diagnostic assessment method for diagnosis assessment, including the Diagnostic Interview for Genetic Study (DIGS) and Medical Chart Records, (3) Using data of impaired attention and executive function for defining endophenotype of SCH in 300 families with at least two siblings affected with schizophrenia. These conditions make the design of this POCOS a break through in current molecular genetic study of SCH nowadays.

The hypotheses to be tested are: (1) There are 3 to 5 vulnerability genes responsible for phenomenological schizophrenia, defined by DSM-IV criteria, locating in chromosome 1q42, 6p22, 8p14, 15q14, 22q12 near markers DIS251, D6S296, D8S1222, D15S976, D22S278, respectively; (2) One vulnerability gene is responsible for endophenotype of schizophrenia, defined by impaired sustained attention assessed by continuous performance test (CPT); This responsible vulnerability gene may be located at chromosome 15q14 near D15S976; (3) Different phenomenological subtypes of schizophrenia, including negative subtype, and non-negative subtype may have different lod scores in linkage analysis with specific genetic markers proposed in this study; (4) There are mutations and/or polymorphisms in the introns and/or exons of the candidate genes associated with the occurrence and/or specific subtypes of SCH.

• Background and Significance. Schizophrenia (SCH) is as devastating and stigmatized psychiatric disorder with brain pathology and high genetic loading. The patients usually become dependent on the family with great social cost. To locate and identify the vulnerability genes, to solve the social stigma, to design genetic counseling can bring a revolutionary development in psychiatry.

This Genomic Study on Schizophrenia（GEMS）comprises of two complimentary component projects to do the positional cloning study on SCH (POCOS) and to do the psychological and genetic counseling study (POGES) in the same pool of families recruited for study. This component project ( NO.1), the POCOS is designed using a breakthrough approach to locate and identify vulnerability genes. The project of POGES (NO.2) is a humanity study complimentary to molecular genetic study and it is designed for exploring psychological issues related to stigma and genetic counseling of this devastating disease of Human Being.

• Genetic Basis of Schizophrenia Genetic epidemiological studies revealed that SCH is familial and the risk to first-degree relatives is approximately ten times the risk to relatives of controls (Tsuang et al., 1980, 1995; Guze et al., 1983; Kendler 1988). Monozygous twin pairs had concordance rates of 46%~53% and 14%~15% for dizygous twin pairs (Kendler KS 1983; Gottesman II 1993; Prescott and Gottesman II, 1993). The heritability was around 0.7. However, the concordance rate in MZ twins is far from 100%, the environmental factors should also be considered. The evidence of genetic contribution to the etiology of SCHwas further supported by adoption study (Heston 1966; Kety et al., 1968,1994; Kendler et al., 1994).

Segregation analyses indicate that the model of multiple genes better fit the observed patterns of schizophrenia in family studies than do single major locus model (Faraone and Tsuang, 1985; Risch and Baron, 1984; Vogler et al., 1990). It was suggested that several genes (3 to 5 in number) in epistasis might responsible for genetic etiology of schizophrenia (Risch, 1990).

• Molecular Genetic Studies of Schizophrenia For linkage analysis, except the sex chromosomes (Delisi and Crow, 1989), there is no a priori hypothesis to focus on any given chromosomal region. The whole genome needs to be systemically screened. A few genome-wide scans of SCH for the decade found that many chromosome regions had suggestive evidences for linkage, including chromosome 1q21-q22, 1q31-q42, 2p22-q21, 4q24-q32, 6p24-p22, 6q16-q23, 8p24-p21, 10p14-p13, 13q14-q32, 15q13-q14, 22q11-q13 (Coon et al., 1994; Shaw et al., 1998; Levinson et al.1998; Blouin et al., 1998; Kaufmann et al.1998; Faraone et al., 1998; Rees et al., 1999; Williams et al., 1999; Hovatta et al., 1999; Brzustowicz et al., 2000). However, only a few chromosome regions were ever reported to have genome-wide significant linkage evidences, including chromosome 1q21-q22 (Brzustowicz et al.,, 2000), 6p24-p22 (Wang et al., 1995), 8p21 (Blouin et al., 1998), and 13q32 (Blouin et al., 1998).

Another promising chromosome region is chromosome 1q 42. A balance translocation (1; 11)(q42.1; q14.3) was associated with major mental illness including schizophrenia in a Scottish large family pedigree (St Clair et al., 1990). Two novel genes named DISC1 (Disrupted in Schizophrenia 1) and DISC2 (Disrupted in Schizophrenia 2) at chromosome 1q42.1 were disrupted at the breakpoint (Millar et al., 2000; 2001). This was confirmed in a Finnish family sample (Hovatta et al., 1999) and another study (Ekelund et al., 2001; Hwu et al, 2001).

All these studies have shown a replication and non-replication pattern (Riley 2000). For detection of genes of modest effect in complex disorders, inadequate sample size and mixed ethnicity were major methodological problems. It is argued at least 600 hundred affected sib-pairs may be required for adequate power (Hauser et al., 1996).

Candidate genes studies revealed inconsistent results in the past decade. Neurotransmitter related genes, such as dopamine (D1, D2, D3, D4, D5), serotonin, r-aminobutyric acid and Glutamate receptor genes had been studied using both association and linkage studies and no consistent results obtained (Asherson et al., 1995, Breyler et al, 1995; Hranilovic et al. 2000; Catalano et al., 1993, Serretti et al., 1999; Chen ACH, 1996,1997). Neuron growth related genes (Margolis et al., 1994), phospholipase genes (Peet, 1998, Wei 1998), and a potassium channel gene (hKCa3/KCNN3) (Dror et al., 1999) have been reported association with schizophrenia. In case-control design, many genes and phenotypes being evaluated and mixture of ethnicity in the sample may inflate the type I errors.

Positional candidate gene approach using linkage dysequilibrium strategy may resolve the above two problems raised by the approach of candidate gene association study basing upon the previous linkage results to increase the prior probability and using parent-offspring trios as internal control. This approach is more powerful than linkage study to locate the susceptibility genes of complex disorder as schizophrenia (Risch and Merikangas, 1996). With the fine mapping linkage evidences, the whole genome sequence and single nucleotide polymorphism (SNP) map, and the advancing microarray technique available, this approach is more efficient to locate the susceptibility genes of schizophrenia, (Owen et al., 2000; Baron 2001). Recently, a study using above strategy has been reported significant linkage dysequilibrium evidence of schizophrenia to a microsatellite polymorphism and a SNP of a gene, NOTCH4 gene, at chromosome 6p21 (Wei and Hemmings, 2000).

Considering the importance of adequate power for linkage analysis and the potentials of positional candidate gene approach using linkage dysequilibrium strategies, we propose this project to do positional cloning of vulnerability genes of schizophrenia. In this stage, we have collected the DNA sample of around 700 families with at least two schizophrenic siblings by our own efforts as well as through collaboration with Harvard University in these four years, and this family sample assures adequate power for linkage analysis and further positional cloning strategy.

• Neuropsychological Deficit in Schizophrenia SCH was fond to have impairment in neuropsychological functions of executive function, sustained attention and working memory (Goldberg and Gold, 1995) and was due to frontostriatal dysfunction (Elliot et al., 1995). Multiple impairment may be best demonstrated by test batteries, rather than single, isolated test (Kremen et al., 1994) andthe impairments strongly suggested dysfunction of frontal-temporal-limbic circuit (Gold and Harvey, 1993).

Visual sustained attention by the Continuous Performance Test (CPT) and executive function by the Wisconsin Card Sorting Test (WCST) were studied more thoroughly. The more difficult ones are stable vulnerability indicators, while the simpler ones might be mediating vulnerability indicators in schizophrenia (Chen and Faraone, 2000). CPT deficits were associated with negative symptoms (Nuechterlein et al., 1986; Hain et al., 1993; Johnstone and Frith, 1996; Liu et al., 1997) and with thought disorder (Nuechterlein et al., 1986; Strauss et al., 1993; Nelson et al., 1998) or disorganized symptoms (Liu et al., 1997).

Deficits in WCST performance were enduring and predicted long term disability, independent of other cognitive deficits (Weinberger et al., 1986; Goldberg et al., 1988). WCST deficits were found to be related to dorsolateral prefrontal cortex (Weinberger et al., 1986; Berman et al., 1995) and that the dopamimetic drugs improves its performance (Daniel et al., 1991; Mattay et al., 1996).

These deficits, being found to be specific to SCH and with genetic risk of SCH, can thus serve as endophenotypes in genetic analysis on SCH.

• Endophenotype Approach in Molecular Genetic Studies of Schizophrenia To resolve the insufficient power of analyses and genetic heterogeneity of SCH, an alternative strategy was to use of a specific neurobiological characteristic of the illness as an endophenotype reflecting the effect of a single genetic alteration (Lander, 1988).

The CPT deficit was a potential endophenotype of the genetic susceptibility to SCH (Chen and Faraone, 2000). It was present not only in SCH patients, but also in their non-psychotic relatives (Grove et al., 1991; Mirsky et al., 1995; Chen et al., 1998). Using data from 148 non-psychotic relatives and 345 community adults, Chen et al. (1998) found that the recurrence risk ratio λwas greater than 15 for the undegraded CPT and greater than 30 for the degraded CPT.

Thus, using CPT deficits as endophenotypes of SCH would provide a valuable measure of genetic risk, would improve the power of genetic analyses and may help identify susceptibility genes for schizophrenia. In our sample, around 220 families have received CPT and WCST assessment. We intent to add 80 families with CPT and WCST data, and to make a 300 of families with available data for endophenotype study. It is feasible to use these endophenotypes for further genetic analysis.

This endophenotype strategy has been successful in mapping of a neurophysiological deficit of schizophrenia, decrease of P50 inhibition, to loci at chromosome 15q13-14, recently. The genome-wide linkage analysis of the P50 inhibition deficit in nine multiplex SCH families found a significant lod score (Z = 5.30,  = 0) at a loci chromosome 15q14. When the clinical diagnosis of SCH was used as the affected phenotype, the maximum lod score at the same marker was not statistically significant (Freedman et al., 1997). The other neurobiological deficit, eye-tracking dysfunction of schizophrenia has been mapped to chromosome 6p23-21 with the maximum multipoint lod score of 4.02. Again, while the clinical diagnosis of schizophrenia was used as the affected phenotype, the linkage result was non-significant (Arolt et al., 1996). In summary, with the endophenotype approach using sustained attention deficits and the adequate power our sample provides, we have confidences in the breakthrough of the searching for vulnerability genes of SCH.

Preliminary Studies.

• Collection of Schizophrenia Co-affected Sibpairs Family The P.I. has been committed himself in collecting the schizophrenia co-affected sib-pairs family since 1990. With the awareness of the importance and critical necessity of diagnostic assessment, these probands, co-affected sib and available non-affected sib as well as the parents were assessed with a semi-structured psychiatrist diagnostic interview (Hwu 1991a) using diagnostic criteria of DSM-III-R and/or DSM-IV. Besides, under the evolutionary theoretical model of psychopathology (Hwu 1985, 1992), developmental data were also collected. All these clinical data and family-tree data were established in the data bank of the molecular genetic laboratory (DBMGL) in the Department of Psychiatry, College of Medicine, National Taiwan University under the auspice of the P.I. In total, there are around 120 schizophrenia co-affected sib-pairs families available for linkage analysis in the DBMGL. As a rule, the DNA samples were obtained from the peripheral white cells of all available subjects of the family, especially the co-affected sibs, at least one non-affected sib and the parents. In total, around 1000 DNA samples were in the DBMGL. Cell-lines of EBV-958 transformed lymphoblast cells were established too. All these study subjects were well informed for this study and informed consent obtained. All families participated in this study were invited to join a " New-Hope Family Club" for periodic meeting and discussion. Around 150 schizophrenia cases who received detail clinical assessment and regular follow-up and comprehensive neuropsychological assessment, including CPT, WCST, WAIS-R, and WMS, were recruited for obtaining DNA samples. Around 200 normal control subjects were also recruited for obtaining DNA samples.

A 4-year nation-wide collaborative work with Harvard University in Taiwan (Taiwan Schizophrenia Linkage Study, TSLS), sponsored by the NIMH, U.S.A., to collect families with co-affected sib-pairs with schizophrenia has been completed. 560 families have been recruited. A total of 600 families will be recruited in this year. DNA samples, cell-lines and clinical data of DIGS, FIGS were collected. Around 220 families received neuropsychological evaluation. (CPT and WCST).

• Molecular Genetic Studies of Schizophrenia Our molecular genetic studies of schizophrenia were supported by three consecutive projects: the molecular genetic project sponsored by the National Science Council (1989-1992), the molecular genetic project of MPGRP (1993-1998) and the molecular genetic project of MPSS (1998-2001) sponsored by NHRI. The early phase (1989-1992) of this molecular genetic project focused on establishing laboratory facilities and collecting co-affected schizophrenic sib-pair families. The 2nd phase of this molecular genetic project (1993-1998) continued the collection of families, and the collection was extended to collect the co-affected bipolar sib-pairs, schizophrenic cases and normal controls. We found the polymorphism of androgen receptor gene of (CAG)n had a probable association with schizophrenia (Hwu et al., 1995). This finding supported the finding of DeLisi et al (1994). The molecular genetic methods of linkage analysis and candidate gene association were promoted, this laboratory of the P.I. (Dr. Hwu) moved to emphasize in this area too. The results of this laboratory were: (1) The polymorphism of (48bp) repeats in DRD4 receptor gene was not significantly linked with schizophrenia (Hong et al., 1998); (2) A single mutation in DRD2 was found not to be associated with schizophrenia (Chen et al.,1996) (3) The samples of the data bank of this project had joined three international collaboration studies using positional cloning approach which need relatively large sample. The one is organized by Gill et al (1996) entitled as "Schizophrenia Collaborative Linkage Group", another was led by Dr. Moises (1995) in Kiel University, Germany and the third one is with Dr. Powell in London (Lin et al., 1995). All these results revealed the possible markers in chromosomes 6p, 11q, 13q, 19q and 22q. This suggests that Taiwanese patients may have possible susceptibility genes in these regions, except chromosome 13q (Lin et al., 1995), fitting an oligogenetic model. (4) The association study on 5-HT2 receptor gene located on chromosome 13q was found to be negative; (5) Clinical epidemiological analysis using co-affected sib-pairs demonstrated the tendency of 3 independent symptom clusters of reality disorganization, disorganization and negative state (Hwu et al., 1997); (6) Weak linkage evidence to loci at chromosome 6p24-22 (Hwu et al.,2000).

The 3rd phase of the molecular genetic study (1998-2001) continued focusing at collecting co-affected schizophrenic sib-pair families and linkage analysis on reported suggestive evidences of chromosome regions, including chromosome 1q21-q22, 1q31-q42, 6p21, 8p24-p21, 15q13-q14, 22q11-q14. The linkage results were (1) suggestive evidence of linkage for schizophrenia with and without the negative symptoms on chromosome 6p24 and 22q12 (Lin et al., 1999a) (2) no linkage evidence of GABAA receptor α1 (GABRA1), β1 (GABRB1) andβ3 (GABAB3) subunit gene with schizophrenia (Lin et al., 1999b) (3) no linkage evidence of Glutamate GluR5 and GluR6 receptor gene with schizophrenia (Lin et al., 1999c) (4) no linkage evidence of SCA1 gene with schizophrenia (Liu et al., 2001a) (5) suggestive linkage evidence on marker (D8s1222) of chromosome 8p with schizophrenia (NPL Z score = 2.58, p=0.005) (Hwu et al., 2001a) (6) suggestive linkage evidence of marker (D1s251) of chromosome 1q31-42 with schizophrenia (NPL Z score = 2.18, p=0.01) (Hwu et al., 2001b), The marker is located near the DISC1 candidate gene. (7) suggestive linkage evidence of marker (D15s976) on 15q13-14 with schizophrenia (NPL Z score = 3.33, p=0.0003) (Liu et al., 2001b). (8) no linkage evidence of schizophrenia to loci at chromosome 1q21-22 (Liu et al., 2001c). (9) Suggestive linkage to NOTCH4 gene locus at chromosome 6p21.3 (NPL Z score = 2.79, p=0.002) (Hwu et al. 2001c). The candidate gene approach has revealed the following results: (1) possible association between Dopamine D4 receptor (DRD4) gene polymorphism with quick treatment response of schizophrenia (Liu et al., 2001d). (2) no association between cytosolic phospholipase A2 (c-PLA2) gene polymorphism and schizophrenia (Liu et al., 2000).

• Studies on Neuropsychological Deficits in Schizophrenia We found impaired sustained attention by continuous performance test (CPT) as the trait marker of schizophrenia. Family studies have indicated that sustained attention deficits as measured by the CPT are vulnerability markers of schizophrenia (Chen and Faraone, 2000). The results are: (1) a substantial proportion of non-psychotic relatives of schizophrenia probands (19-34%) have CPT deficits, which can be predicted from their probands' CPT performance (Chen et al., 1998a); (2) subjects with schizotypal personality features also exhibit CPT deficits, which are specifically associated with negative factors of schizotypy. (Chen et al., 1998b); (3) CPT deficits are present in schizophrenic patients, are particularly associated with negative and disorganized symptoms, and those with more difficult CPT versions are not amenable to neuroleptic treatment (Liu et al., 2000).

The specificity of CPT was studied (Liu et al., 2000) in a group of schizophrenia patients (n=41) in contrast to the group of bipolar patients with psychotic symptoms (n=46) and the group of bipolar patients without psychotic symptoms (n=22) and a group of patients with non-psychotic major depressive disorder (n=22). It was found that CPT deficits are stable vulnerable indicators of schizophrenia, mediating vulnerability indicators for bipolar disorder, and state-dependent indicator for major depression. These results demonstrate that CPT deficits are valid trait marker of schizophrenia.

• Studies on Clinical Heterogeneity of Schizophrenia In the field of psychiatry, clinical psychopathological background is crucial for molecular genetic studies. The principal investigator (HGH) had developed an evolutionary psychopathological theoretical model for descriptive as well as neurobiological studies. The complete work of the descriptive psychopathological study on schizophrenia was published [Schizophrenia: a descriptive psychopathology, 1999 ISBN 957-9201-21-8, Taipei, Chu-ching Publishing Company]. The clinical heterogeneity issue has been approached by a novel statistical method using graphic plotting technology (Chen, 1999) based on PANSS rating scale. Two (Hwu et al, 2001) or three (Hwu et al, 2001) subtypes of schizophrenia were delineated and validated by follow-up and neuropsychological data such as attention impairment (CPT). Through two-year follow-up, the symptom patterns of schizophrenic patients were stable (Hwu et al, 2001). This confirms the subtyping efforts in schizophrenia using phenomenological symptom patterns. These data suggest the possible fruitfulness of genetic linkage study on schizophrenia considering the phenomenological subtypes as well as endophenotypes in linkage as well as association studies.

In conclusion, all these preliminary data reveal that the DNA sample, clinical and endophenotype data have been well prepared in this POCOS program. The PI and his team are experienced in performing the laboratory work and further genetic analysis in this project.

Inclusion criteria

• Clinical diagnosis of Schizophrenia
• Two children and one parent in the same family all have the clinical diagnosis of schizophrenia

Exclusion criteria

• One parent should be normal