Healthy Context: Genomic Imprinting

Genomic imprinting restricts the expression of nearly 150 mammalian genes to a single parental allele. These genes play a key role in both growth control and the brain. In addition to cancers, deregulation of imprinting leads to several rare pathologies, including some severe neurodevelopmental disorders, highlighting the importance of defining the tissue-specific control mechanisms of imprinting in order to establish the causes of these pathologies.

At the molecular level, these genes are grouped into about twenty domains, each regulated by an Imprinting Control Region (ICR) and the concerted action of several levels of epigenetic regulation. However, the mechanisms that coordinate this multi-scale regulation and allow the ICR to remotely instruct the mono-allelic expression of domain genes during the acquisition of a cellular identity remain poorly understood.

Our data indicate the importance of bivalence dynamics, and of the H3K27me3 mark in particular, at the ICR in this process. They also suggest the involvement of the 3D genome structure and chromatin loops through which the ICR can physically interact with distal genes and regulatory regions.

Based on these findings, we have developed an integrated multi-scale approach combining molecular, epigenetic, 3D genome conformation and bioinformatics analyses conducted in a corticogenesis model to identify the factors that coordinate the expression of imprinted genes during murine neural stem cell formation.

Pathological Context : Glioma

This project aims to identify the mechanisms that coordinate the different levels of epigenetic regulation by investigating the causes and consequences of their alteration in malignant glioma, one of the most common types of brain tumours. Collaborations with clinical teams allow us to carry out this study on samples from the Auvergne Glioma Tumorotheque and glioblastoma initiating cell lines (GIC). An exhaustive molecular study allowed us to reveal that the majority of transcriptional defects in gliomas are not related to DNA methylation defects but rather due to an alteration in the control of H3K27me3 (here). The search for the causes of this alteration leads us to study in more detail the (de)regulation of HOX genes, for which we have established that the alteration is a signature of the most aggressive gliomas.

To identify new candidates we have undertaken, in a parallel approach, to explore the link between deregulation of transposable elements and alteration of H3K27me3 in cancer cells. Our original hypothesis is that antisense transcription from L1 repeats can induce aberrant expression of adjacent genes, in the form of so-called LINE-1 chimeric transcripts (LCTs), which in turn influence the dynamics of the H3K27me3 mark. We have previously developed the CLIFinder bioinformatics tool to identify these chimeric transcripts from RNA seq data (here). With this tool, we have, for the first time, established the panorama of LCTs in aggressive gliomas and undertaken to characterise the altered genes in CIGs.