miRNAs and networks of co-expressed genes associated with childhood trauma: from molecular scars to drug discovery in bipolar disorders – CT-ScarBip

The clinical validation of childhood maltreatment as a predictor of mood recurrences in BD is based on a large cohort of patients with an extensive clinical characterization. We will use survival analyses to demonstrate that childhood maltreatment increases the risk of a subsequent mood recurrence during the follow-up of two years.
The starting point of the molecular part of the project is the generation of a blood whole-genome mRNA and miRNA signature that is associated with childhood maltreatment.
The first branch includes the integration of mRNA and miRNA sequencing data, the validation of identified genes and miRNAs in independent samples of patients and the construction of networks of co-expressed genes and of the interactions between DEG and DEmiR. We will construct co-expression networks using the WGCNA algorithm to define modules of co-expression. For each module we will test functional enrichment. We will then apply a drug prioritization approach based on the « signature reversion paradigm ». First, using a large collection of drug-induced gene expression data (Connectivity Map), drugs will be prioritized based on their ability to restore the regulation of modules identified. Prioritized drugs will be validated in cell lines of patients with BD to demonstrate their ability to reverse the signature as compared to lithium and valproate.
The second branch consists in a cross-species approach that aims at providing evidence for overlaps between molecular signatures in human blood versus mice blood and then mice brain. One of the outputs of this step is to provide a limited set of DemiR that will be manipulated in mice to observe behavioural consequences.

The clinical evaluation is carried out at inclusion and the exposure to childhood maltreatment is performed using the CTQ (Childhood Trauma questionnaire) which allows an evaluation of five subtypes of trauma (emotional and physical neglect, emotional abuse, physical and sexual). We were able to show in a sample of 1908 patients that exposure to childhood maltreatment is associated with an earlier relapse during follow-up, which is particularly significant with physical abuse (HR = 1.05 [1.02-1.08]; p = 0.0014).
The analyses from RNA and miRNA sequencing are on-going.

Our project should allow us to identify blood-based CT-related markers and to propose a more tailored drug prescription, aiming at decreasing recurrences rate in BD. In addition, the identification of this CT-related signature will help us to characterize and better understand the mechanisms at stake in patients exposed to early life stress.

Delahaye-Duriez A, Mokhtari A, Etain B, Yalcin I, Marie-Claire C, Ibrahim E, Belzeaux R, Lutz P-E. Transcriptome analysis to identify co-expressed gene networks as a molecular signature for childhood trauma-related mood disorders. Poster presented at the congress of the American Society of Human Genetics october2019 in Houston, Texas, USA

Mokhtari A, Etain B, Yalcin I, Marie-Claire C, Ibrahim E, Belzeaux R, Lutz P-E and Delahaye-Duriez A. Transcriptome analysis to identify co-expressed gene networks as a molecular signature for childhood trauma-related mood disorders. Poster presented at the JOBIM (Journées Ouvertes de Biologie, Informatique et Mathématiques) july 2019 in Nantes, France

Réda C, Kaufmann E, and Delahaye-Duriez A. Machine learning applications in drug development. Comput. Struct. Biotechnol. J. 2019 doi:10.1016/j.csbj.2019.12.006. [Epub ahead of print] hal.inria.fr/hal-02533303/

Bipolar disorder (BD) is a frequent, severe and burdensome psychiatric disorder. The severity of BD mainly relies on the risk of mood recurrences that are major providers of suicidal risk, hospitalizations, cognitive decline and psychosocial consequences. Predicting mood recurrences is thus a challenge for targeting patients who would require a more intense care plan (patient stratification) and moving towards personalized medicine in BD (drug prioritization).
We have demonstrated that Childhood Traumatic events (CT) dramatically increase the severity of the clinical expression of BD and decrease the response to lithium, the gold standard mood stabilizer used in BD. Importantly, CT have been repeatedly found to increase the frequency of mood recurrences, but only in retrospective studies. Altogether, these findings highlight that CT may act as an environmental “disease modifier” in BD.
The biological consequences of CT are still poorly understood. As evidenced in preclinical models, CT in humans have long lasting effects, both in brain and in blood, on the expression of genes involved in multiple biological pathways, such as systems related to stress reactivity (cortisol/HPA axis), immunity/inflammation, neurotransmission or neuroplasticity. Within mechanisms implicated in the regulation of transcriptional processes, microRNAs (miRNAs) are proposed as key players in epigenetic regulation that might play a role in adaptive mechanisms linked to the response to stress. Integrating different levels of alterations in gene expression (mRNA) and transcriptional regulation (miRNA) might provide a more comprehensive physiopathological pathway leading from CT exposure to recurrences in BD.
Finally, further characterization of genes whose expression is modified in association with the exposure to CT might open avenues in the identification of new therapeutic targets. Such integrative genomic approaches that combine high throughput sequencing of mRNAs and miRNAs represent unique tools for drug repurposing and prioritization.
The first clinical objective of this project is a validation of CT as a disease modifier (prediction of mood recurrences) in a prospective sample of patients with BD, most previous studies using a retrospective design.
The second molecular objective is to generate an integrated mRNA and miRNA blood signature that is associated with the exposure to CT in patients with BD. This signature will be validated in independent samples of patients and healthy controls according to their levels of CT. This signature will be functionally validated both in silico and in vitro (interactions between specific mRNAs and miRNAs).
The third therapeutic objective is to use this signature for drug repurposing, meaning that the identified networks of co-expressed genes will serve to the identification and prioritization of already known drugs that would be able to modulate the gene expression modifications associated with CT. The effects of these drugs to restore the molecular scars associated with CT will be tested on patients’ cells and compared to the effects of gold standards mood stabilizers used in BD, such as lithium and valproate.
The fourth pathophysiological objective is to use a cross-species approach to test for the overlap or congruence of early life stress signatures observed in patients (mRNA/miRNA modifications in blood) and mice (mRNA/miRNA modifications in blood and brain). This will serve to identify miRNAs whose expression will be down-regulated in specific brain regions in mice to characterize the behavioral consequences.
The project will therefore help patients stratification, provide a mRNA and miRNA signature in the blood of patients that will be used for drug repurposing and prioritization (improving thus personalized medicine) and allow further understanding in the complex molecular abnormalities that represent scars of CT.