The Maudsley Family Study of Psychosis

A Quest for Intermediate Phenotypes

Colm McDonald

• Publisher: Psychology Press
• Year: 2008
• Print ISBN: 9781841697345
• eBook ISBN: 9781135430092
• Edition: 1
• Topic: Psychologie
• Subtopic: Kognitive Psychologie & Kognition

Psychotic disorders such as schizophrenia and bipolar disorder are known to be highly heritable. Despite decades of research, however, the genetic variations conferring susceptibility to these illnesses have yet to be identified. Such genetic variations most likely produce abnormalities of brain structure and function from which the clinical features of psychosis emerge.

The Maudsley Family Study of Psychosis investigates the genetically produced markers of abnormal brain structure and function ('intermediate phenotypes') which underlie the clinical syndrome of schizophrenia, and more recently bipolar disorder. In this book, key findings of this important research program, and their implications for this field, are discussed in detail. Contributors outline research examining brain structure and functioning in patients with schizophrenia and their unaffected first degree relatives, incorporating detailed clinical assessments, magnetic resonance imaging, electrophysiology, eye tracking measures and neuropsychology.

This book provides an improved understanding of illness pathways and potential scope for intervention in order to better manage and prevent psychotic disorders and will be of interest to academics and clinicians in the field.

CHAPTER ONE
Exploring intermediate phenotypes of psychosis

Colm McDonald, Robin Murray

THE CONTRIBUTION OF GENETIC LIABILITY TO SCHIZOPHRENIA

The tendency for schizophrenia to run in families has long been recognized. A century ago, Emil Kraepelin, who described the syndrome of ‘dementia praecox’, the forerunner of schizophrenia, noted: ‘I had myself found formerly in Heidelberg general hereditary predisposition to dementia praecox in about 70 per cent of the cases in which about this point reliable statements were to hand’ (Kraepelin, 1919). Since Kraepelin’s initial observations, overwhelming evidence has emerged from family, twin and adoption studies that schizophrenia has a strong genetic component.

Family studies

In a seminal study pooling the results of about forty European family studies performed between 1920 and 1987, Gottesman (1991) demonstrated that the lifetime morbid risk of developing schizophrenia among the relatives of patients with schizophrenia increased with the degree of genetic relatedness to the affected individual. The risk to third-degree relatives was 2%, to second-degree relatives around 4–6%, to siblings or children around 9–13% and the risk to identical twins or the offspring of dual matings was 46–48%. Furthermore, the risk increased if more than one relative was affected, for example, the risk if both a sibling and parent are affected was 17%. The majority of the studies summarized by Gottesman (1991) were performed before the advent of operationalized criteria for diagnosing schizophrenia. However, several studies that did employ operationalized diagnostic criteria have essentially confirmed the results of these earlier studies (Gershon et al., 1988; Kendler et al., 1993a; Maier et al., 1993).

Twin studies

Although family studies show clearly that schizophrenia is familial, they do not confirm that this relationship is genetic, as family members also share much of their environment in common. The relative contributions of genetic and environmental risk factors can be disentangled by studying rates of concordance for the disorder in twins. If both members of a twin pair have schizophrenia, they are classified as concordant, whereas if only one member of a twin pair is affected, they are classified as discordant. A disorder is likely to be under genetic influence if concordance rates are higher in monozygotic (MZ) twins, who share 100% of their genes, than dizygotic (DZ) twins, who share 50% of their genes on average. The results of all the major twin studies across several different countries find the concordance rate for MZ twins to be substantially higher than that found for DZ twins. Kendler (1983) pooled the results of twelve such studies and found the probandwise concordance rate for MZ twins to be 53%, whereas the rate for DZ twins is 15%. Again, more recent studies using operationalized diagnostic criteria for schizophrenia confirmed the markedly higher MZ than DZ probandwise concordance rates (McGuffin et al., 1984; Onstad et al., 1991). Although these studies provide substantial evidence for a genetic contribution to schizophrenia, the lack of 100% concordance among monozygotic twins also provides compelling support for the importance of environmental factors in contributing to schizophrenia.

Adoption studies

The relative contributions of genetic and environmental risk factors to the aetiology of schizophrenia can also be dissected by the use of adoption studies. In addition to shared genes, relatives (and in particular identical twins) also share an extensive environment in common, including social and cultural behaviours, biological hazards and psychological stresses, which could include common risk factors for schizophrenia. A series of naturalistic adoption studies have been performed in an attempt to minimize the effect of such common environmental risk factors in family studies of schizophrenia by comparing rates in biological and rates in non-biological relatives of patients with schizophrenia. These studies have found higher rates of schizophrenia in the adopted-away offspring of mothers of patients with schizophrenia compared to adopted-away offspring of controls (Heston, 1966; Rosenthal et al., 1975) and even compared to the rate in a small group of subjects who were adopted by parents one of whom later developed schizophrenia (Wender et al., 1974).

Furthermore, the rates of schizophrenia or schizophrenia-spectrum disorders among the biological relatives of adoptees with schizophrenia is higher than the rates in adoptive relatives and the relatives of control adoptees (Kety, 1983; Kety et al., 1994), including a higher rate of schizophrenia in paternal half siblings of adoptees with schizophrenia than in paternal half siblings of adoptees without schizophrenia, indicating that high rates of schizophrenia in the offspring of patients with schizophrenia are not related to prenatal or perinatal influences (Kety, 1988). As with family and twin studies, the results of these early adoption studies have been confirmed when more rigorous operationally defined diagnostic criteria are used to diagnose schizophrenia and its related disorders (Kendler et al., 1994; Tienari et al., 1994). Thus, the combined findings from these adoption studies provide important support for the conclusions, derived from family and twin studies, that familial clustering of schizophrenia is an expression of shared genetic factors rather than shared environmental factors.

HERITABILITY ESTIMATES FOR SCHIZOPHRENIA

Statistical models can be applied to data derived from twin studies to estimate the likely heritability of an illness—the proportion of the variance in liability contributed to by genes. Estimates of heritability vary across samples and methods of ascertainment but usually involve a model that includes estimates of genetic effects, common environmental effects and non-shared environmental effects. Such estimates have ranged between 41% and 87% for schizophrenia (Cardno et al., 1999; Kendler, 1983), with the heritability estimates using operationally defined diagnoses tending to be in the upper end of this range (Farmer et al., 1987; Onstad et al., 1991). It has also been reported that the common environmental component can be removed from the model without weakening the fit but increasing the heritability to 87%, with the remainder of the liability explained by non-shared environmental effects (Cardno et al., 1999; McGuffin et al., 1994). In further support of these findings, Cannon et al. (1998) used structural equation modelling in a population cohort of Finnish twins, thus excluding any bias associated with estimates of liability based on index twins, and demonstrated that 83% of the variance in liability was due to additive genetic factors, with the remaining 17% due to unique environmental factors.

MODE OF GENETIC TRANSMISSION

Although results from family, twin and adoption studies provide evidence that genetic factors play a part in the aetiology of schizophrenia, the exact mechanisms of genetic transmission remain unidentified. The simplest model of genetic transmission is one in which a single gene is responsible for the illness. If such a model were true, genetic phenomena such as incomplete penetrance (where genetic variation is not always expressed in the clinical phenotype) and pleiotropy (where a single genetic alteration can result in variable phenotypic expression) are required to explain the apparent non-Mendelian pattern of transmission and heterogenous clinical presentation of these illnesses. However, studies on the recurrence risk from twin and family studies using statistical modelling have demonstrated that schizophrenia is very unlikely to be the product of a single gene or a collection of single gene disorders, even taking into account incomplete penetrance (McGue et al., 1985; O’Rourke et al., 1982).

More complex models of genetic transmission involving multiple genes and environmental risk factors are more likely to be responsible for the patterns of inheritance observed. One such model is the liability/threshold model, which was first applied to schizophrenia by Gottesman and Shields (1967). In this model, the liability to develop schizophrenia is normally distributed in the population and is due to multiple genes of small effect acting additively and in combination with environmental risk factors, but only those individuals whose liability exceeds a certain critical threshold manifest the illness. Relatives have an increased liability compared to the general population, and thus a higher proportion of relatives also lie beyond the threshold (reflected in the higher prevalence of schizophrenia in the relatives of patients). The number of genes involved in such a model is unpredictable and inheritance could be oligogenic—a small number of genes of moderate effect (e.g. fewer than ten) or polygenic—multiple genes of small effect (e.g. more than 100).

MOLECULAR GENETICS

Linkage studies

Linkage occurs when a genetic marker and a disease gene lie close to each other on the same chromosome. In this case, the marker and the disease gene will be found to occur together more often in affected family members than would be expected by chance. Linkage studies require families that contain several affected members and are most appropriately employed to detect a small number of genes of relatively large effect, i.e. mono/oligogenic inheritance rather than polygenic inheritance. Historically, the results of linkage studies in schizophrenia were disappointing and characterized by multiple failed replications, most likely contributed to by a combination of weak genetic effects and small sample sizes. In recent years, there has been some progress with successful replications of linkage to several chromosomal regions. This was facilitated by the development of highly polymorphic genetic markers evenly spaced throughout the genome, which enabled genomewide scans for susceptibility loci, and by large-scale international collaborations to achieve greater statistical power through combining large numbers of subjects. Two meta-analyses in recent years (with different methodologies) of genomewide schizophrenia linkage studies have supported the existence of susceptibility genes on chromosomes 8p and 22q (Badner & Gershon, 2002; Lewis et al., 2003) with further strong support for loci at 13q (Badner & Gershon, 2002) and 2q (Lewis et al., 2003), and weaker support for loci at 1q, 3p, 5q, 6p, 11q, 14p and 20q (Lewis et al., 2003). More recent genomewide linkage studies in various populations continue to emerge at a rapid pace and implicate other chromosomal regions including 2q37 (Wijsman et al., 2003), 10q22 (Faraone et al., 2006), 11p (Suarez et al., 2006) and 18pll (Faraone et al., 2005). Some of the chromosomal loci reported may well prove to be false positives in due course, but others are likely to harbour susceptibility genes.

Association studies

Association studies require no major assumption other than the existence of a genetic contribution to the disorder. In contrast to linkage analysis, the aim is to examine the frequency of marker alleles in a sample of unrelated patients compared to a sample of ethnically matched controls. Association studies are more appropriately employed in the detection of polygenic inheritance, in which a large number of genes have relatively minor effects. A higher frequency of the marker allele in the patient group suggests that the marker allele is itself related to susceptibility to the disease or else is closely linked to the disease allele. A disadvantage is that the marker must be very tightly linked to the disease gene (<1 cM). Consequently, such studies have largely been confined to testing candidate genes thought to have functional significance in the illness or to fine mapping of chromosomal loci previously identified as showing linkage. There have been many negative studies and failed replications using the former technique, but some support has emerged for weak associations between schizophrenia and polymorphisms of certain genes controlling neurotransmission pathways, including the dopamine receptor DRD2 (Glatt et al., 2003b), dopamine receptor DRD3 (Jonsson et al., 2003), serotonergic receptor 5-HT2A (Abdolmaleky et al., 2004), metabotropic glutamate receptor GRM3 (Chen et al., 2005; Egan et al., 2004) a...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• List of contributors
• List of figures
• List of colour plates
• List of tables
• Preface
• 1. Exploring intermediate phenotypes of psychosis
• 2. The Maudsley Family Study of Psychosis—overview of clinical methodology and characteristics
• 3. Auditory evoked potentials as genetic trait markers of schizophrenia
• 4. Are eye movement abnormalities related to susceptibility genes for schizophrenia?
• 5. Neuropsychological impairments in patients with schizophrenia and their unaffected relatives
• 6. Neurological abnormalities in patients with schizophrenia from singly and multiply affected families and their relatives
• 7. Structural brain deviations in schizophrenia and bipolar disorder—to what extent are they genetically mediated?
• 8. Summary and implications
• Appendix 1