Establishment of a schizophrenia classifier based on peripheral blood signatures and investigation of pathogenic miRNA-mRNA regulation

To date, the diagnosis of schizophrenia (SCZ) mainly relies on patients' or guardians' self-reports and clinical observation, and the pathogenesis of SCZ remains elusive. In this study, we sought to develop a reliable classifier for diagnosing SCZ patients and provide clues to the etiology and pathogenesis of SCZ. Based on the high throughput sequencing analysis of peripheral blood miRNA expression profile and weighted gene co-expression network analysis (WGCNA) in our previous study, we selected eleven hub miRNAs for validation by qRT-PCR in 51 SCZ patients and 51 controls. miR-939-5p, miR-4732-3p let-7d-3p, and miR-142-3p were confirmed to be significantly up-regulated, and miR-30e-3p and miR-23a-3p were down-regulated in SCZ patients. miR-30e-3p with the most considerable fold change and statistically significance was selected for targeting validation. We first performed bioinformatics prediction followed by qRT-PCR and verified the up-regulation of potential target mRNAs (ABI1, NMT1, HMGB1) expression. Next, we found that the expression level of ABI1 was significantly up-regulated in SH-SY5Y cells transfected with miR-30e-3p mimics. Lastly, we conducted a luciferase assay in 293T cells confirming that miR-30e-3p could directly bind with the 3'untranslated region (3'-UTR) of ABI1, revealing that miR-30e-3p might play a role in the polymerization of neuronal actin and the reconstruction of the cytoskeleton via the downstream regulation of ABI1. In addition, we constructed a classifier by a series of bioinformatics algorithms and evaluated its diagnostic performance. It appears that the classifier consists of miRNAs and mRNAs possess a better discrimination performance than individual miRNA or mRNA in SCZ.