Identification of serum microRNAs as diagnostic biomarkers for schizophrenia

Background At present, the schizophrenia diagnoses are based on the clinical symptoms and behaviors neglecting the laboratory test indicators. Results To better investigate the diagnostic potential of miRNAs for schizophrenia, we selected 14 candidate miRNAs and examined their expressions in the serums of 40 schizophrenia patients and 40 healthy controls by qRT-PCR. Ultimately three abnormally expressed microRNAs were identified, i.e., miR-34a-5p, miR-432-5p and miR-449a. Then, binary regression analysis was employed to combine 3 dysregulated miRNAs. ROC analysis revealed that the AUC of the combination of miR-432-5p + miR-449a in serums was 0.841 (95% CI: 0.791~0.887) with 90% sensitivity and 80% specificity. The AUC of the combination of miR-34a-5p + miR-432-5p + miR-449a in serums was 0.843 (95% CI: 0.791~0.887) with 90% sensitivity and 77.5% specificity. The results indicated that the combined model of miR-432-5p + miR-449a and miR-34a-5p + miR-432-5p + miR-449a have better prediction performances. Conclusions The study concludes that the two miRNAs combinations have the potential to be used as biomarkers for schizophrenia diagnoses. The finding may be conducive to overcoming the dilemmas faced by current schizophrenia diagnosis.


Background
Schizophrenia is one of the common kinds of mental diseases that has a lifetime risk of nearly 1% [1] . Currently, schizophrenia diagnoses are based on the clinical assessment of symptoms and behavior abnormalities of patients. The laboratory tests of pathophysiological and biochemical indicators are not available to assist the clinical diagnoses of schizophrenia. MiRNAs are a class of short endogenous non-coding RNA, which participate in the post-transcriptional regulation of protein-coding genes [2,3]. It has been found that miRNAs regulate many essential biological functions and processes [4,5] and are involved in almost all the life processes [6]. The changes of miRNA expressions reflect the alteration of neuropsychiatric disorders in the genetic and biological aspects [7,8], including schizophrenia [9][10][11].

Participants
All the subjects of the study were diagnosed as schizophrenia patients by at least two experienced psychiatrists based on the criteria specified in the Diagnosis and Statistical Manual of Mental Disorders Fourth Edition (DSM-IV) and International Classification of Diseases 10 (ICD10). All the patients were treated with antipsychotic drugs including aripiprazole, risperidone, olanzapine, and clozapine. Control groups were recruited from volunteers. Blood donors and any individual who suffered or are suffering a mental illness and whose relatives suffered or are suffering with mental illness were excluded. Before conducting this study, we obtained the written consent from all the participants. The study was approved by the ethical committee of Tongliao Institute of Mental Health, Tongliao, China.

Peripheral blood collection
Independent peripheral blood samples (5 ml per participant) was obtained from 80 subjects in the morning on an empty stomach and were collected in EDTA anticoagulant tubes. After standing for half an hour on ice, the supernatant was aspirated. 10,000 g Centrifuged for 5 min and discard the precipitate. The supernatant that left was stored at − 80°C until using for extraction of total RNA.

Statistical analyses
Data were analyzed using SPSS software version 20

Discussion
The current study selected 14 candidate miRNAs and examined their expressions in the serums to explore their diagnostic value for schizophrenia. The qRT-PCR result showed that the expressions of miR-34a-5p, miR-432-5p, and miR-449a have difference in schizophrenia patients. The AUC of the combinations of miR-432-5p + miR-449a and miR-34a-5p + miR432-5p + miR-449a were greater compared with any single miRNA, indicating that the two combination of miRNAs can be used as biomarkers for schizophrenia. The study confirmed that miR-34a, miR-432, and miR-449a expressed significant changes in serums of schizophrenia patients, suggesting that they may play an important role in the development of schizophrenia. MiR-34a has some well-documented involvements in neurogenesis and neural differentiation [28,29]. The change of the expression of miR-34a has been found in brain tissues of patients with psychiatric disorder [30,31]. Schipper et al. (2007) found that the expression of miR-34a significantly up-regulated in the PBMCs of patients with Alzheimer's disease [32]. The expression level of miR-432 was observed to alter in postmortem cerebellar cortex from autism patients [33]. The miR-449 expression level was found deregulated in cerebrospinal fluid of AD patient brains [34]. Wu et al. (2014) verified that miR-449 and miR-34b/c are very important in normal brain development [35]. Using the results of genome-wide association analysis (GWAS) of schizophrenia, Hauberg et al. predict miRNAs target genes and performed regression and enrichment analysis to explore the regulation of miRNAs on schizophrenia risk genes, they listed 10 conserved miRNAs in the enrichment analysis, including miR-34 ac-5p and miR-449a, and indicated these miRNAs play an important role in schizophrenia [36]. In addition, several target genes of hsa-miR-34a were found using the predictions of Tar-getScan, such as GREM2, TANC2, CAMSAP1, RGMB, CALN1, RTN4RL1 and FKBP1B are related to the development and function of neuron [16]. The study showed that miR-34a-5p, miR-432-5p, and miR-449a expressed aberrantly in the serums of schizophrenia patients. Before then,  verified the abnormal expression of miR-34a existed in PBMCs of schizophrenia patients [19].  found again the expressions of miR-34a and miR-432 exhibited the difference in plasma [18]. Yu et al. (2014) verified that the expression level of miR-432 were significantly down-regulated  in PBMCs of schizophrenia patients before treatment compared with healthy controls [23]. Lai et al. (2016) verified that the miR-34a, miR-432 and miR-449a did not change in the PBMCs of patients that hospitalized for more than two months and whose symptoms were alleviated [16]. They believe the hospitalization and symptom alleviation could not affected the expressions of circulating miRNAs, so they were sufficiently stable and detectable biomarkers in peripheral blood [16]. In addition, Lai et al. (2014) investigated the expression changes of blood-based miRNA from preterm infants to adulthood, and found miR-34a, miR-432, and miR-449a were expressed consistently from infancy to adulthood [37]. The limitations of the study includes the lack of distinction among psychotic symptoms and the limited number of subjects. All the selected patients in the current study were treated with antipsychotic drugs. We can not exclude the possibility of the expressions of change in these miRNAs before and after the treatment.