Impact of genetic and environmental risk factors for schizophrenia on cortical brain structure
It is now accepted that Schizophrenia, a neurodevelopmental disorder which affects around 1% of the population, is influenced by both genetic and environmental risk factors. Schizophrenia is evidenced as being heritable with twin-heritability estimates of around 80%. Recently, the disorder has been demonstrated to be polygenic in nature; many genetic variants with individually small effects contribute to the overall phenotypic variation. Furthermore, cannabis use, adverse events, urbanicity, obstetric complications and migration, are five environmental risk factors that have been evidenced as being associated with an increased risk of developing the disorder. Abnormalities in brain structure are also well evidenced in individuals with schizophrenia, in particular, reduced cortical thickness, volume and surface area have been linked to those with schizophrenia when compared to healthy controls. It has been posited that these cortical alterations may predate disorder onset, for example, disruptions in brain development may be a function of experiencing schizophrenia-associated genetic and environmental risk factors. However, the link between genes, environment and brain structure within schizophrenia remains unclear. In this thesis, we aimed to examine whether genetic and environmental risk factors for schizophrenia directly impact cortical brain structure. Methods and Results The current aims were assessed using measures of cortical thickness, volume and surface area, as defined by FreeSurfer, in three separate studies. Firstly, ANCOVA models were applied to a case-control sample, the Scottish Family Mental Health (SFMH) study, ncontrols/npatients = 41/58) to determine whether PolyGenic Risk Scores for Schizophrenia (PGRS-SCZ) are associated with lower cortical thickness both globally and within regions of interest (frontal and temporal lobes) as well as to examine whether the effects of experiencing an accumulation of the five environmental risk factors (outlined above) is associated with greater cortical thinning (Chapter Two). The results indicated that an increased PGRS-SCZ was related to lower, global cortical thickness in the whole sample and not a result of group differences. With regards to environmental effects, the more environmental risk factors experienced, the lower the cortical thickness, this was specific to the temporal lobe. Secondly, to further investigate the link between environmental risk factors of schizophrenia, we focused on birth weight as a proxy for obstetric complications (Chapter Three). Linear mixed effects regression (LME) models were used to assess whether birth weight was associated with cortical thickness, surface area and volume in a UK Biobank (UKB) sample (n = 1,680). We then applied Mendelian Randomisation (MR) to determine if birth weight-associated genetic variants were causally related to cortical structure. The results in this chapter suggested that higher weight at birth was associated with larger cortical volumes and surface area, both globally and in several cortical sub-regions. In contrast, a negative association was found between birth weight and cortical thickness in the lateral occipital parcel. MR analysis suggested a causal link of birth weight, as indexed by genetic variants, and insular lobe cortical volume as well as surface area globally, in the insular lobe and in middle temporal, medial orbitofrontal and inferior frontal gyrus parcels. Lastly, we tested whether the same association between PGRS-SCZ and cortical thickness (outlined in Chapter Two) could be replicated within a subset of UKB (Chapter Four). For this, we again utilised LME models using the second genetic data release of UKB (n = 2,864). We tested this globally, lobarly and within 27 bilateral cortical parcels for each of these parameters. We found a higher PGRS-SCZ to be associated with lower global cortical volume and thickness as well as insular lobe cortical thickness. To further test potential environmental influences (as outlined in Chapter Three) on these effects, we used a liner regression model to test for a relationship between PGRS-SCZ and birth weight as well as LME models to test for interactional effects. No relationship was found between PGRS-SCZ and birth weight nor were there any significant interactions found between PGRS-SCZ and birth weight on cortical structure Conclusion Together, these studies highlight the fact that both genetic and environmental risk factors for schizophrenia may, indeed, directly but differentially impact cortical brain structure. This information may help us to further understand the progression of the disorder but also, by identifying and addressing these risk factors early, we may be able to minimise the impact that the disorder can have on cortical brain structure; particularly in relation to potentially modifiable factors, such as birth weight. We also highlight the importance of using large samples and replications in order to examine such relationships.