From risk genes to psychiatric phenotypes - Studies of fibroblast growth factor-related and genome-wide genetic variants in humans and mice

Schizophrenia is a severe mental disorder with a high heritability. This thesis describes studies on the association between genetic variants and phenotypes related to schizophrenia, such as brain volume and IQ, in order to learn about which processes are affected by schizophrenia-associated genetic variants. The first three chapters focus on fibroblast growth factors (FGFs) as candidate genes for schizophrenia. In Chapter 2, the functions of FGFs in neurogenesis, axon outgrowth and repair after stress and their associations with psychiatric disorders are reviewed. Mutant mice for some FGFs and its receptors showed neuronal and behavioral abnormalities similar to those seen in psychiatric patients, such as reduced brain volume and impairments in cognition. Chapter 3 reports that common genetic variants (single nucleotide polymorphisms or SNPs) in the FGF system are not associated with brain volume in schizophrenia patients and in healthy control subjects. In Chapter 4, we report a knockout mice study that investigates the effects of deletion of FGF2 and its interaction with early life stress. Mice lacking FGF2 showed increased locomotor activity under baseline conditions and increased aggression towards same-sex intruder mice. Interestingly, early life stress normalized the effect of FGF2 depletion on hyperactivity, perhaps due to increased sheltering behavior. Aggressiveness of FGF2-/- mice was not altered by early life stress. In addition, early life stress was associated with reduced neuronal differentiation in the hippocampus and FGF2-/- mice were shown to be especially sensitive to this. In the next two chapters, we investigated the association between a polygenic schizophrenia risk score (PSS) and brain volume and IQ. In Chapter 5, we describe that the PSS was significantly associated with total brain volume and white matter volume, equally in schizophrenia patients and healthy controls. A group of 186 SNPs showed most evidence for association with schizophrenia and white matter volume and an explorative functional analysis showed that these SNPs were located in genes with neuronal functions. Remarkably, the PSS was neither associated with IQ within the patient group, nor within the healthy control group, as described in Chapter 6. The last three chapters focus on copy number variations (CNVs), relatively large stretches of deleted or duplicated DNA. Global CNV burden was not associated with IQ, nor with brain volume, as described in Chapter 6 and 7, respectively. In Chapter 8, a study on the association between CNVs and paternal age in schizophrenia patients is described. Schizophrenia patients have on average older fathers than healthy controls. We report that subjects with older fathers (35+ years at the time of birth) carry around 2 times more rare, large deletions compared to subjects with younger fathers, while there was no change in the numbers of other types of CNVs. The effect was observed for both schizophrenia patients and healthy controls. These rare, large deletions have a higher chance to be derived de novo than the average CNV. Therefore, a possible mechanism is that de novo mutations accumulate in the sperm of older males.