CORDIS - EU research results

Investigating epigenetic silencing mechanisms within cancer

Final Report Summary - BET-HUIDOBRO (Investigating epigenetic silencing mechanisms within cancer)

The heterochromatin refers to highly compacted regions of chromatin, which in general harbours transcriptionally repressed regions of the DNA. This conformation is dynamic, as in facultative heterochromatin where structural domains can change from an open to a closed conformation, or vice-versa, to coordinate gene expression during development. The mechanisms underlying this conformational changes, or whether they are cause or consequence of transcription, remains to be elucidated.
The experimental approach is based on the use of stable neocentromeres, centromeres that appear at new locations in the genome that were previously considered as non-centromeric. Centromeres are cytogenetically defined as the primary constriction where the binding of the two sister chromatids takes place, after DNA replication. Epigenetic mechanisms are actively involved in the maintenance of this chromosome feature. In humans, the centromere core domain contains CENP-A (H3 histone variant) nucleosomes that are wrapped around repetitive DNA (alpha satellite repeats). Flanking this core domain, pericentromeric heterochromatin is often found, playing an important role for centromere function. Both the centromere and the pericentromeric regions are defined by epigenetic mechanisms.
Active transcription from the centromere core domain is required to maintain centromere identity and function in yeast, but whether this also happens in mammalian cells remains currently unknown. Previous works have found transcription at centromeres in human cancer cell lines. This transcription seems to occur at specific stages of the cell cycle (mitosis) and gives rise to small transcripts that contribute to the maintenance of the centromere.
Therefore, neocentromeres appear to be a much more robust model to investigate the relation between transcriptional activity and chromatin architecture. To carry out the project, patient-derived cells were used in parallel to human-hamster cell lines carrying genetically isolated human chromosomes having a neocentromere.
Objective 1: Chemical screening and chromatin characterization
Firstly, chromatin immunoprecipitation using centromeric proteins was performed in combination with DNA arrays (ChIP-on-chip) to accurately map the neocentromere to a 200 Mb region in the long arm of the chromosome 3. Neocentromeres were found to originate in regions which lack of the alpha-satellite DNA repeats commonly found in canonical human centromeres. Instead, they were located in the context of non-repetitive DNA, and deep sequencing could not detect differences between chromosome 3 containing the centromere or not.
Histone proteins are found in cell nuclei, where they bind to DNA to form nucleosomes, the elementary unit of DNA packaging in eukaryotes. Histone N-terminal tails protrude from the nucleosomes and are susceptible of a varied range of chemical modifications which together with other factors, regulate the accessibility of DNA to the transcription machinery.
Histone marks profiling was also performed by ChIP-on-chip. Marks that are related with an open configuration of the DNA did not show any changes comparing the neocentrome with the equivalent region in the normal chromosome. However, these marks were depleted from actively transcribed genes in the flanking regions of the neocentromere. Our region of interest exhibited enrichment in a histone mark that targets transcriptionally silent and compacted chromatin, which inversely correlated with DNA methylation. Endogenous centromeres exhibit high levels of histone 3 lysine 9 methylation at the neighbouring regions, which were also found at the flanking regions of our neocentromere.
Pre-eliminary experiments in the showed that although the neocentromere is seen as a primary constriction (equal to canonical centromeres), chromatin within the neocentromere is in a more relaxed conformation than in the corresponding region in normal chromosomes. To highlight, in the presence of transcription inhibitors, the difference in compaction was restored to the same extent when the neocentromere was absent, suggesting that transcription could be closely related to chromatin architecture at neocentromeres.
To monitor transcription at the neocentromere, binding of RNA polymerases II (RNAPII) and III (RNAPIII) was investigated. RNA polymerases are the enzymes that produce RNA using DNA as template. RNAPIII is absent at neocentromeres, but RNAPII binding is slightly increased in the cell line carrying the neocentromere, corroborating previous data obtained in the lab which revealed RNAPII presence at centromeres by immunofluorescence. As a novelty, this binding occurs on interphase cells, suggesting that centromeric transcription occurs not only during mitosis.
Transcription was directly assessed by RNA sequencing (RNA-seq). Despite no transcripts have been found at the neocentromeric region, the neighbouring genes which lose active histone marks were transcriptionally inactive, corroborating previous data. Centromeric transcripts are expected to be small and highly unstable, and they are probably quickly degraded by the exosome complex, the major RNA degradation machinery present in eukaryotic cells. Unsuccessful attempts were made to deplete cells from the exosome complex, whose activity may be hindering detection of transcription at the neocentromeres. We are currently working on this but if this approach is unsuccessful capture-RNA-seq may be used to maximise read depth to characterise transcription at the centromeres
Objective 2: To investigate the consequences of the ectopic recruitment of the BET domain proteins to genomic DNA loci using RNA-guided targeting technology
Since histone acetylation (recognised by the BET domain) did not seem to have any functional relevance for the neocentromere, it was decided to ectopically recruit other candidate proteins and transcription factors to the neocentromeres. This would allow us to directly assess the impact of transcription on chromatin structure and vice-versa. Both RNA (CRISPR/Cas9) and non-RNA (TALENs) guided techniques technology have been initiated to perturb chromatin the environment and the transcriptional state at the neocentromere to investigate the consequences. Covadonga is currently building up the tools to recruit transcriptional activators (VP64) and repressors (Sid4X).
In this project, the epigenetic patterns of a naturally occurring and a fully functional human neocentromere have been thoroughly characterized providing a better understanding about how this essential region of chromosomes is defined. As a novelty, changes in chromatin architecture within the context of centromeres have been identify, and we provide evidence suggesting that centromeric transcription is playing a major role in this process. The artificial manipulation of the neocentromere could provide a better understanding of how centromeres work. We hope that genetic engineer experiments will cause re-activation of the original centromere or induce chromosome lost.