Schizophrenia

From DNA to behaviour – Imaging genetics and epigenetics in Schizophrenia

Approximately 0.7 % of the world population suffers from schizophrenia (van Os and Kapur 2009) and a combination of genetic and environmental factors are responsible for developing this disease. The heritability is estimated at 80% (Gottesman and Gould 2003) and first-degree relatives of schizophrenia patients have the highest risk of being affected (Gottesman 1991). Research results have already linked alterations in the dopamine- and glutamate-system to the emergence and course of this disease. However, the exact biochemical aberrations remain unknown. This knowledge would be an important basis for the development of better pharmacological agents and a more targeted treatment strategy.

Genetic association studies can help to understand the underlying biological mechanisms. Several genetic risk variants like NRG1 and DISC1 have repeatedly been linked to schizophrenia (Duff et al. 2013; Ruano et al. 2008; Ohi et al. 2012) and with the increasing sample sizes of psychiatric consortia the discovery of replicable genetic loci associated with schizophrenia is now feasible (Ripke et al. 2013). However, research findings from conventional genetic association studies have shown inconsistencies concerning heritable mental disorders like schizophrenia (O’Donovan, Craddock, and Owen 2009). These results may be caused by ill-defined clinical phenotypes in psychiatry. Hence scientific efforts have increased to investigate the association between genetic polymorphisms and so-called intermediate phenotypes (Tan, Callicott, and Weinberger 2008; Gottesman and Gould 2003) which are thought to be more proximal to the underlying substrate of the illness since they are independent of diagnosis and statistically more powerful.

Among the most well-known heritable intermediate phenotypes of schizophrenia are prefrontal inefficiency during working memory tasks, reduced hippocampal volume and reduced cortical thickness (Ehrlich et al. 2011; Hall and Smoller 2010). Our group and other scientists were already able to verify associations between candidate genes and some of these phenotypes (Walton et al. 2013a; Brauns et al. 2013; Brauns et al. 2011; Ehrlich et al. 2010). Nevertheless, the exploration of new biological mechanisms may be impeded as the candidate gene approach is based on known pathophysiological premises of schizophrenia. Since well-established genetic variants only account for a fraction of the variance, we aim to analyze genome-wide data in combination with intermediate phenotypes (Walton et al. 2013b; Hass et al. 2013).

To fully entangle the cause of mental disorders like schizophrenia, environmental factors such as the individual living environment, drug abuse and prenatal stressors have to be incorporated into the mathematical models. First steps have been made by analyzing methylome-wide data which – to a certain degree – can reflect the influence of environmental factors at the epigenetic level (Gavin and Sharma 2010; Roth et al. 2009; van Os, Rutten, and Poulton 2008). However the correlation between DNA methylation measured and blood and methylation measured in brain tissue is rather low (Walton et al., 2015).

Our group uses clinical, neurophysiological, genome-wide SNP and methylome-wide CpG as well as multi-modal imaging data from large-scale multicenter studies on schizophrenia (MCIC, ENIGMA, GENUS) to test new concepts of data reduction and multivariate data analysis (Walton et al. 2013b). We aim to identify associations between (poly)genetic and (poly)epigenetic risk-factors and biomarkers of schizophrenia at the same time be investigated brain correlates of clinical symptoms (Walton et al., 2017a; Walton et al., 2017b).