Abstract
BACKGROUND: Schizophrenia has been associated with deficits across numerous cognitive domains. As cognitive impairments are poorly treated by antipsychotics, it is imperative to understand the molecular mechanisms influencing these deficits. While candidate gene studies have extensively explored cognitive performance in schizophrenia, there is a lack of studies investigating gene expression. As schizophrenia is a complex clinical disease, using gene expression profiling gives detailed insight to potential pathways involved in cognitive impairment. Research has shown that cognitive deficits should be investigated by separate domains rather than global cognition. As one of the most replicated deficits in schizophrenia is in working memory, the study aimed to determine the changes in gene expression associated with working memory deficit in schizophrenia. METHODS: Thirty-six patients with a DSM-IV diagnosis of schizophrenia completed the Digit Span Backwards and Letter Number Span (LNS) tasks as a measure of working memory. The Digit Span backwards raw score and LNS total raw score were used as variables in a k-means cluster analysis to divide patients into those that showed a working memory (WM) deficit compared to those that had intact WM. mRNA was extracted from peripheral blood samples donated by all patients. mRNA levels between the WM deficit and intact WM groups were compared using the AffymetrixTM Human Exon 1.0 ST Array. Genes with significant changes were analyzed using Ingenuity Pathway Analysis to identify the pathways that most impacted by changes in gene expression and differentiated the two groups. RESULTS: The WM deficit group were 17 patients with schizophrenia (10 male) and the intact WM group were 19 patients (11 male). There was a significant difference between the two groups in both WM tasks (p < 0.001), but no significant difference in age, years of education, premorbid IQ, symptom severity (PANSS) or duration of illness. There was a significant variation in mRNA levels between the two groups in 164 genes. IPA identified the following as top pathways that differentiated the WM deficit and intact WM groups: NRG1 interactome (8 genes identified; p = 1.70x10-2), Estrogen 2 group (9 genes identified; p = 2.94x10-2), COMT (2 genes identified; p = 4.35x10-2) and Glutamate (9 genes identified; p = 7.35x10-2) pathways. Possible interactomes of proven gene x gene interactions were also constructed using the genes with most significant gene expression differences. DISCUSSION: Overall, our data suggests that working memory deficits can be attributed to an interplay between glutamatergic, dopaminergic and estrogen classical pathways. This is consistent with previous literature indicating dopamine-glutamate interactions influencing cognition in schizophrenia and is also consistent with previous research providing evidence for protective effects of estradiol on dopamine and glutamate systems. As the study measured gene expression in peripheral blood samples, the results should be interpreted with caution and further research should be conducted. However, the consistency of the current findings with previous literature support the use of peripheral blood as a possible representation of expression in brain tissue.