ABSTRACT
The study of patients with schizophrenia (SZ) at different clinical stages may help clarify what effects could be due to the disease itself, to the pharmacological treatment, or to the disease progression. We compared expression levels of targeted genes in blood from individuals in different stages of SZ: clinical high risk for psychosis (CHR), first episode of psychosis (FEP), and chronic SZ (CSZ). Then, we further verified whether single-nucleotide polymorphisms (SNPs) could be related to gene expression differences. We investigated 12 genes in 394 individuals (27 individuals with CHR, 70 antipsychotic-naive individuals with FEP, 157 CSZ patients, and 140 healthy controls (HCs)). For a subsample, genotype data were also available, and we extracted SNPs that were previously associated with the expression of selected genes in whole blood or brain tissue. We generated a mediation model in which a putative cause (SNP) is related to a presumed effect (disorder) via an intermediate variable (gene expression). MBP and NDEL1 were upregulated in FEP compared to all other groups; DGCR8 was downregulated in FEP compared to HC and CHR; DGCR2 was downregulated in CSZ compared to FEP and HCs; DISC1 was upregulated in schizophrenia compared to controls or FEP, possibly induced by the rs3738398 and rs10864693 genotypes, which were associated with DISC1 expression; and UFD1 was upregulated in CSZ and CHR compared to FEP and HC. Our results indicated changes in gene expression profiles throughout the different clinical stages of SZ, reinforcing the need for staging approaches to better capture SZ heterogeneity.
ABSTRACT
In this study, we aimed to test if the schizophrenia (SCZ) polygenic risk score (PRS) was associated with clinical symptoms in (a) the first episode of psychosis pre-treatment (FEP), (b) at nine weeks after initiation of risperidone treatment (FEP-9W) and (c) with the response to risperidone. We performed a detailed clinical assessment of 60 FEP patients who were antipsychotic-naive and, again, after nine weeks of standardized treatment with risperidone. After blood collection and DNA isolation, the samples were genotyped using the Illumina PsychArrayChip and then imputed. To calculate PRS, we used the latest available GWAS summary statistics from the Psychiatric Genomics Consortium wave-2 SCZ group as a training set. We used Poisson regression to test association between PRS and clinical measurements correcting for the four principal components (genotyping). We considered a p-value < 0.0014 (Bonferroni correction) as significant. First, we verified that the schizophrenia PRS was also able to distinguish cases from controls in this south-eastern Brazilian sample, with a similar variance explained to that seen in Northern European populations. In addition, within-cases analyses, we found that PRS is significantly correlated with baseline (pre-treatment) symptoms, as measured by lower clinical global assessment of functioning (-GAF), higher depressive symptoms and higher scores on a derived excitement factor. After standardized treatment for nine weeks, the correlation with GAF and the excitement factor disappeared while depressive symptoms became negatively associated with PRS. We conclude that drug (and other treatments) may confound attempts to understand the aetiological influence on symptomatology of polygenic risk scores. These results highlight the importance of studying schizophrenia, and other disorders, pre-treatment to understand the relationship between polygenic risk and phenotypic features.