A game changer for bipolar disorder diagnosis using RNA editing-based biomarkers

Abstract

In clinical practice, differentiating Bipolar Disorder (BD) from unipolar depression is a challenge due to the depressive symptoms, which are the core presentations of both disorders. This misdiagnosis during depressive episodes results in a delay in proper treatment and a poor management of their condition. In a first step, using A-to-I RNA editome analysis, we discovered 646 variants (366 genes) differentially edited between depressed patients and healthy volunteers in a discovery cohort of 57 participants. After using stringent criteria and biological pathway analysis, candidate biomarkers from 8 genes were singled out and tested in a validation cohort of 410 participants. Combining the selected biomarkers with a machine learning approach achieved to discriminate depressed patients (n = 267) versus controls (n = 143) with an AUC of 0.930 (CI 95% [0.879-0.982]), a sensitivity of 84.0% and a specificity of 87.1%. In a second step by selecting among the depressed patients those with unipolar depression (n = 160) or BD (n = 95), we identified a combination of 6 biomarkers which allowed a differential diagnosis of bipolar disorder with an AUC of 0.935 and high specificity (Sp = 84.6%) and sensitivity (Se = 90.9%). The association of RNA editing variants modifications with depression subtypes and the use of artificial intelligence allowed developing a new tool to identify, among depressed patients, those suffering from BD. This test will help to reduce the misdiagnosis delay of bipolar patients, leading to an earlier implementation of a proper treatment.

Conflict of interest statement

The authors declare no competing interests.

© 2022. The Author(s).

Figures

Fig. 1. STARD flow chart diagram of participants included in the study.

The diagram reports the flow of participants through the study, indicating the number of participants evaluated for the study, the number of participants excluded because they did not meet the inclusion criteria, and the number of participants included in the discovery and in the validation study.

Fig. 2. RNA editing landscape in whole blood.

A Proportions of RNA variant types in human whole blood. A-to-G variant, indicating A-to-I editing, is disproportionately enriched. B The percentage of RNA editing sites across different intervals of editing degrees. C The proportion and numbers of A-to-I editing sites across genomic localization are shown in piechart and at the right of the gray bars. The proportion of editing sites not residing in Alu repeats is represented by black bars. D Repartition of A-to-I editing events or edited genes by chromosome or gene. Black (right y-axis): Repartition of A-to-I editing events by chromosome. Red (left y-axis): Number of edited genes by chromosome. E Distribution of Alu editing index (AEI) values in controls and depressed patients. Data are the mean ± SEM. p value of AEI index was calculated using the Wilcoxon rank-sum test.

Fig. 3. Identification of differentially A-to-I RNA edited sites between healthy controls and depressed patients using RNA-Seq data from human blood.

A Volcano plot of differentially edited sites between healthy controls and depressed patients. The volcano plot shows the upregulated and downregulated sites differentially edited between depressed patients healthy controls. For each plot, the x-axis represents the log2(Fold Change) (FC), and the y-axis represents -log10(p values). Editing sites with a p value of < 0.05 were assigned as differentially edited and are indicated in green. Significant editing sites of selected genes (see Table 2) are labeled in black. B: Heatmap of MeSH class associated with the 7 genes identified by RNAseq annotated by the DisGeNET database. All identified genes have been analyzed and their relationships with mental disorders, behavior and behavior mechanisms, nervous system diseases and immune system diseases MeSH categories have been calculated with DisGeNET database. The darker the heatmap, the stronger the association.

Fig. 4. Diagnostic performance of the tests.

Target Editing Index (TEI) were calculated by combining all significant RNA editing variants with p value ≤ 0.05 for each target gene. ROC curve plotted are the probabilities for the correct response of the tested sets using case specific trained random forest models. The results of the 100 trained random forest by MultDS were combined by majority voting for test dataset. RF results are shown on the test dataset which has never seen the algorithm. The implementation was done using the randomForest and caret R package. A TEI for depressed patients (DEP; n = 267) and healthy controls (CTRL; n = 143); (B) ROC curve and diagnostic performance of Random Forest model for DEP vs CTRL classification; (C) TEI for unipolar (n = 160) and bipolar disorder (BD; n = 95); (D) ROC curve and diagnostic performance of Random Forest model for unipolar vs BD classification.