A Study on Risk Mutations of Vulnerability Genes of Schizophrenia

4a. Specific Aims. In our previous work based on a Taiwanese sample, we have found that schizophrenia was significantly associated with 11 candidate genes, including GNPAT (1q42.1), DISC1 (1q42.1), MRDS1 (6p24.3), NOTCH4 (6p21.3), NRG1 (8p21-p12), DAAO (12q22), G72 (13q32), CHRNA7 (15q14), PRODH (22q11.21), HMOXI (22q12), and CACNG2 (22q12). In this 3-year proposal, we have 5 specific aims, in which aim 1 is a continuation of our previous study of the Positional Cloning Study of SCH (POCOS) using family data, while aims 2 to 5 are newly developed research goals.

Specific Aim 1: To investigate tighter association of these 11 candidate vulnerability genes with different subtypes defined by clinical or endophenotype variables. The endophenotype variables include the neuropsychological functions of sustained attention measured by continuous performance test (CPT), executive function measured by Wisconsin Card Sorting Test (WSCT) and the skin physiological function of niacin flushing responses.

The hypothesis to be tested is that schizophrenia is genetically heterogeneous and the specific risk SNP markers or specific risk SNPs haplotypes of different candidate genes will be associated with specific clinical subtype or specific endophenotype indicators with higher degree of statistical significance.

Specific Aim 2: To search for functional risk mutations of these candidate genes using direct sequencing technique, and to investigate the contributing genetic variance of the specific risk mutations in specific subtypes of schizophrenia, as mentioned in specific Aim 1, using case-control study design. The interaction effects of these functional risk mutations will be explored.

The hypothesis to be tested is that there are several specific risk mutations from 11 candidate genes. The contributing variances of these different risk mutations will be different.

Specific Aim 3: To find the variation of functional expressions of these identified risk mutations in Epstein-Barr virus (EBV) transformed lymphoblastoid cell lines of the patient subjects, and neuronal cell-lines.

We hypothesize that these candidate genes have expression discriminations when carrying risk mutations, which will be able to be detected from EBV-transformed lymphoblastoid cells of these patients and neuronal cell-lines or from constructed reporter gene expression.

Specific Aim 4: To explore and clarify the candidate genes responsible for the impaired niacin skin flushing response in schizophrenia.

We hypothesized that difference in the skin niacin flushing response will reveals difference in the gene expression in the lymphocytes using microarray approach based on a matched case- control design.

Specific Aim 5: To establish a procedure to assess the prepulse inhibition function (PPI) in normal mice as the endophenotype indicator of schizophrenia, and test its validity with acute or chronic administration of amphetamine. This animal model will serve further functional study on the knock-out or knock-down mice carrying these identified risk mutations.

The hypothesis to be tested is that acute or chronic administration of amphetamine will block PPI in wild-type mice.

In addition, we will explore during the conducting of this proposal the following feasibilities: (1) To obtain human brain tissues for studying the functional expressions of these vulnerability genes with specific risk mutations; (2) To employ the animal model to assess neurophysiological and behavioral abnormalities in mice carrying the functional risk mutations; (3) To explore novel treatment for preventing the abnormal neurobiological process due to the risk mutations; and (4)To develop early detection and early intervention program, including genetic counseling, based on the data of risk mutations.

4. Research Plan (continuation) 4b. Background and Significance. Schizophrenia (SCH) is a devastating and stigmatized psychiatric disorder with high genetic loading. The identification of vulnerability genes of SCH is at dawn nowadays. As the vulnerability genes and the specific risk mutations are identified, there will come the revolutionary development in SCH research and clinical practice, including reliable and valid diagnosis based on genetic and related neurobiological data, novel treatment and early prevention of schizophrenia based on the knowledge of abnormal gene function and its interaction with the environmental factors.

The Genomic Psychiatry Study Group (GENOP), a multidisciplinary team of National Taiwan University Hospital, has been successfully operating in studying genetic factors in the psychopathology of SCH. This GENOP has carried out the positional cloning study on vulnerability genes of Schizophrenia (POCOS), supported by NRPGM in these 3 years, and has found 11 candidates vulnerability genes through dense SNP fine mapping study. This GENOP also have a thorough neurobiological and clinical psychopathological study (Hwu et al., 2002; Chen et al., 1998; Chiu et al., 2004; Hsieh et al., 2004; Hwu, 1999; Hwu et al., 2003). Thus, the GENOP is ready to take a step further to search for risk mutations of SCH as proposed here.

We conclude that the finding of vulnerability genes for schizophrenia is at dawn based on the following lines of empirical evidence:

1. Genetic Epidemiological Study SCH is a highly heritable disease, based on family, twin (Kendler KS 1983; Gottesman II 1993; Prescott and Gottesman II, 1993) and adoption (Heston 1966; Kety et al., 1968, 1994; Kendler et al., 1994) studies with an heritability of 0.70 (Tsuang et al., 1980, 1995; Guze et al., 1983; Kendler, 1988) and a λ, risk to first-degree relatives versus controls, value of 10 (Kendler KS, 1983; Gottesman II 1993; Prescott and Gottesman II, 1993; Chang et al., 2002). An oligogenetic model of SCH etiology was commonly accepted (Faraone and Tsuang, 1985; Risch and Baron, 1984; Vogler et al., 1990, Risch, 1990).

2. Molecular Genetic Studies of Schizophrenia 2-1. Linkage Analyses Genome-wide scans of SCH for the decade found that many chromosome regions had suggestive evidence for linkage (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), whereas subsequent studies revealed a replication and non-replication pattern (Riley, 2000). Inadequate sample size and mixed ethnicity might be the major methodological problems (Altmuller et al., 2001, Hauser et al., 1996).

2-2. Candidate chromosome region approach A balance translocation (1; 11) (q42.1; q14.3) was found to be associated with major mental illness including schizophrenia (St Clair et al., 1990). The novel gene DISC1 at chromosome 1q42.1 was repeatedly found to have significant association with SCH, including our Taiwanese samples (Millar et al., 2000; 2001; Hovatta et al., 1999; Ekelund et al., 2001; Hwu et al, 2001; Ekelund et al., 2004, Liu YL et al., 2004).

2-3. Candidate gene approach Many candidate genes, such as dopamine (D1, D2, D3, D4, and D5), serotonin, and glutamate receptor genes, based on psychopharmacological hypothesis (Asherson et al., 1995, Breyler et al, 1995; Hranilovic et al. 2000; Catalano et al., 1993, Serretti et al., 1999; Chen et al., 1996, 1997), as well as developmental- and physiological-based genes such as 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 to have association, though not consistently, with schizophrenia. The mixture of ethnicity in the sample may inflate the type I errors.

2-4. Association study using SNP fine mapping and family internal control method By genetic epidemiological guideline such as using linkage disequilibrium strategy and parent-offspring trios as internal control to guide the improved method to locate the susceptibility genes of complex disorder (Risch and Merikangas, 1996; Owen et al., 2000; Baron 2001), there were about 10 candidate genes found by using this modality of study approach (Elkin et al., 2004). These candidate genes included RGS4 (1q) (Mirnics et al., 2001), MRDS1 (6p24.3), DTNBP1 (6p22.3) (Straub et al., 2002; Tang et al., 2003; van Den Bogaert et al., 2003; van den Oord et al., 2003) , TNFa (6p22) (Boin et al., 2001; Pae et al., 2003), NOTCH4 (6p21.3) (Wei and Hemmings, 2000; Fan et al., 2002; Sklar et al., 2001; Skol et al., 2003; Takahashi et al., 2003), PPP3CC (8p21.3) (Gerber et al., 2003), NRG1(8p21-p12)( Stefansson et al., 2002; 2003;Yang et al., 2003) , DAAO(12q22) (Chumakov et al., 2002), G72(13q33) (Chumakov et al., 2002), CHRNA7(15q14) (Leonard S et al., 2002) , PRODH(22q11.21) (Liu et al., 2002; Jacquet et al., 2002) , COMT(22q11.21) (Shifman et al., 2002). Nevertheless, all these candidate genes still wait to be replicated in different study samples.

The GENOP of NTUH, by using dense SNP fine mapping method, have found that 8 significant candidate genes of MRDS1(6p24.3), NOTCH4(6p21.3), NRG1(8p21-p12), DAAO(12q22), G72(13q32), and PRODH(22q11.21), HMOXI (22q12), and CACNG2 (22q12) (Liu et al., 2004). This research team is thus ready to take a step further to search for the risk mutations of these candidate genes and the validation study of these risk mutations. Based on the meta-analysis of published studies, Altmuller et al (2001) and Risch et al.(2000) suggested that high density SNP association analysis in combination with functional genomic data may be necessary to detect susceptibility loci (which may be of small effect) for complex human disease, as shown in a few recent reports (Emamian et al., 2004; Mukai et al., 2004). This proposed RIGOS will follow this strategy as the main approach as shown in our specific aims.

3. Endophenotype indictors and Issue of Heterogeneity of SCH 3-1. Neurobiological Impairment in Schizophrenia SCH was found to have impairment in neuropsychological functions of executive function, sustained attention measured and working memory (Goldberg et al, 1993) with underlied frontostriatal dysfunction (Elliot et al., 1995), or dysfunction of frontal-temporal-limbic circuit (Gold and Harvey, 1993), or dorsolateral prefrontal cortex (Weinberger et al., 1988; Berman et al., 1988).

These impaired neurobiological functions had significant association with specific clinical symptoms (Nuechterlein et al., 1986; Hain et al., 1993; Johnstone and Frith, 1996; Liu et al., 1997; Nuechterlein et al., 1986; Strauss et al., 1993; Nelson et al., 1998; Liu et al., 1997) and enduring long term disability (Weinberger et al., 1986; Goldberg et al., 1988).

SCH is characterized by core-deficits in information processing and attentional mechanism (Bleuler, 1911; Kraepelin and Robertson, 1919; McGhie and Chapman, 1961; Venables, 1964; Braff and Geyer, 1990), which was due to the impairment of "filtering" or "gating" function of brain (Braff and Geyer 1990). P50 event-related potential suppression (P50I) [Freedman et al., 1987; Olincy et al., 2000] and prepulse inhibition of the acoustic startle reflex (PPI) (Braff et al., 2001; Geyer et al., 2001) are reliable measurements. Both P50I and PPI mechanisms, under different neuronal circuits (Swerdlow et al., 2001; Adler et al., 1998), are out of subject's will control, and both are impaired in the patients of SCH (Freedman et al., 1987; Braff et al., 2001), in the spectrum disorder of schizotypal disorder (Cadenhead et al., 1998; Cadenhead et al., 2000) and clinically unaffected relatives of SCH patients (Siegel et al., 1984; Waldo et al., 1988; Waldo et al., 1991; Adler et al., 1992; Clementz et al., 1998; Cadenhead et al., 2000). Both deficits in P50I and PPI can be as the endophenotype indicators with different neural substrate processes. In this study, we intend to use P50I and PPI as specific endophenotype indicator for identifying the possible responsible risk mutations of the possible candidate vulnerability genes.

Animal studies have shown that the PPI deficits could be induced pharmacologically by agents altering systems implicated in the neurobiology of schizophrenia, such as dopaminergic or NMDA glutamatergic neurotransmission (Mansbach et al., 1988; Hutchison & Swift, 1999). This experimentally induced PPI impairment by amphetamine has been conceived as a putative animal model with face, predictive, and construct validities for certain schizophrenic symptoms (Koch, 1999; Swerdlow et al, 2000).

Impairment of Niacin Skin Flush Response in Schizophrenia: The niacin induced vasodilatory response was either absent or impaired in schizophrenic patients (Horrobin, 1980; Wilson and Douglass, 1986; Fielder et al., 1986; Hudson et al., 1997; Rybakowski and Weterle, 1991; Hudson et al., 1997, 1999; Ward et al., 1998; Puri et al., 2001; Shah et al., 2000; Maclean et al., 2003). There is no significant relationship between antipsychotic drug dose and niacin sensitivity (Hundson et al., 1997; Messamore et al., 2003).

The niacin sensitivity may be through the nicotinic acid receptor HM74 (low affinity receptor) and HM74a (high affinity receptor) recently identified (Wise et al., 2003; Soga et al., 2003), and they are both Gi-G-protein-coupled receptors (GPCR). The two receptors were expressed in the lymphocytes (Wise et al., 2003). The detail molecules and signaling pathway involved in niacin skin flush response still remained to be clarified. In this RIGOS project, it is desirable to explore the possible responsible genes in this un-studied area using microarray method.

3-2. Endophenotype Approach in Molecular Genetic Studies of Schizophrenia To face the genetic heterogeneity of SCH, an alternative strategy was to use specific neurobiological characteristic of the illness as an endophenotype reflecting the effect of a single genetic alteration (Lander, 1988). In this respect, deficits on the Continuous Performance Test (CPT) have been postulated to be a potential endophenotype of the genetic susceptibility to SCH (Chen and Faraone, 2000; Grove et al., 1991; Mirsky et al., 1995; Chen et al., 1998).

CPT deficit was a potential endophenotype of the genetic susceptibility to SCH (Chen and Faraone, 2000; Grove et al., 1991; Mirsky et al., 1995; Chen et al., 1998).

This endophenotype strategy has been successful in mapping of a neurophysiological deficit, decrease of P50 inhibition, to loci at chromosome 15q13-14 (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 (Arolt et al., 1996). This RIGOS project will take this endophenotype strategy for validation of risk mutations, especially by using attention impairment measured by CPT as well as P50I and PPI indicators.

3-3. Issue of Heterogeneity Genetic and clinical heterogeneity is the main issue in this area of study. Searching for valid clinical subtype and endophenotype are critical for a breaking through study. Our GENOP group had substantial psychopathological data in this area and the RIGOS project will take this strategy.

4. Functions of Candidate Vulnerability Genes This RIGOS project will tackle on these 11 candidates vulnerability genes identified by our previous POCOS project. The neuronal functions of these 11 candidate genes to be studied have not been well characterized. Current knowledge on the functions of some of these 11 candidate genes are listed as the following: the GNPAT is a dihydroxyacetonephosphateacyltransferase (DHAPAT, or DAPAT) participating in the etherphospholipid biosynthesis (Ofman et al., 2001), the NOTCH4 is a heterodimer receptor involving in the vascular and articular cartilage development (Hayes et al., 2003; Iso et al., 2003), the NRG1 (neuregulin 1) is a group of isoform proteins produced from NRG1 gene by alternative splicing, these isomers interact with ERBB receptor and induce the growth and differentiation of neuronal and glial cells (Kerber et al., 2003; Fallon et al., 2004). The DAAO (D-amino-acid oxidase) is a peroxisomal enzyme activated by G72 (D-amino acid oxidase activator) (Chumakov et al., 2002). The CHRNA7 is a nicotinic acetylcholine receptor subunit belonging to the superfamily of ligand-gated ion channels and involving in the development of the central nervous system (Agulhon et al., 1999; Raux et al., 2002). The PRODH (proline dehydrogenase) is an enzyme catalyzing proline to Δ-1-pyrroline-5-carboxylate (P5C), the first step in the conversion of proline to glutamate (Harrison and Owen 2003; Lee et al., 2003). Deletion of this gene has been associated with hyperprolinemia (Jacquet et al., 2002). The functional roles of these genes on the neuronal cells warrant further investigation to predict the possible pathological process for schizophrenia. In this RIGOS project, we'll use in vitro cell-line to do functional laboratory study to validate any of these risk mutations identified in this RIGOS project.