X-chromosome inactivation

Philip Avner, Corinne Chureau, Philippe Clerc, Lynda Deuve, Agnès Dubois, Julie Legoupi, Sylvain Maenner, Céline Morey, Sara Merzouk, Julie Prudhomme, Nisa Renault

Male mammals carrying one X and one Y chromosome and females with two X chromosomes could be courting potential disaster due to the imbalance in the number of copies of X-linked genes between the sexes. Nature’s answer to this is to arrange for the transcriptional silencing of all the 1500 odd genes present on one of the two X chromosomes in the female during early embryonic development. This process, known as X-inactivation, is a striking example of epigenetic gene regulation, and requires that the two chromosome homologues are differentially treated within the same nucleus. X-inactivation is under the control of a master control region on the X chromosome, the Xic (X-inactivation centre). The onset or initiation of X-inactivation involves the cell counting how many X-chromosomes are present, then ensuring that only a single X chromosome remains active in the diploid cell. Initiation is also thought to include a recognition process linked to the choice of the X chromosome to be inactivated (e.g. imprinting).
A key player in the Xic is the Xist gene which is transcribed as a non coding RNA. Expressed from each X chromosome before differentiation, the Xist transcript is stabilized and coats the inactive X chromosome once inactivation is initiated (see figure on right).
Once X-inactivation is initiated, the chosen X-chromosome is epigenetically modified, accumulating successively a series of other modifications that often characterise heterochromatin. Changes include modifications of histone proteins, accumulation of polycomb group proteins and CpG DNA methylation.

Xist decoration during X-chromosome

inactivation:

The laboratory’s main interests concern

1) The early regulation of X-inactivation through the genetic dissection of the Xic. Varied approaches including targeted mutagenesis, transgenesis, genomic and transcriptional analysis and biochemical characterisation are being used.

2) The chromatin modifications associated with the accumulation of Xist RNA on the X-chromosome, and which result in transcriptional silencing.

Much of our work in the laboratory depends on exploiting ex vivo model systems for X inactivation - embryonic stem (ES) cells, trophoblastic stem (TS) cells and extraembryonic endoderm (XEN) cells. Female ES cells possess two active X chromosomes, one of which is inactivated upon differentiation. The entire sequence of events that characterizes random X-inactivation in vivo is observed in ES cells. TS and XEN cells show imprinted X-inactivation, similar to that observed in murine extraembryonic tissues.

Dynamics of nuclear re-organisation associated with genome reprogramming in mouse pre-implantation embryos

Jane Deuve, Agnès Dubois, Sara Merzouk, Julie Prudhomme, Céline Morey
Cell pluripotency and genome plasticity at pre-implantation stages are necessary to ensure a proper establishment of the various cell lineages of the embryo. Epigenetic marks such as specific chromatin structures or a specific organisation of the chromatin fibres within the cell nucleus might underlie this plasticity. To characterize these marks, we study the X-chromosome inactivation process, which is considered as a paradigm of large scale epigenetic reprogramming.
Figure 1. X-chromosome inactivation during pre-implantation development.
The mouse morula shows an imprinted inactivation of the paternal X (Xp) (Okamoto et al. 2004). In the early blastocyst - a time when differentiation of the first cell lineages occurs - this imprinted X-inactivation is conserved in extra-embryonic derivatives (primitive endoderm and trophectoderm) but is reverted in the inner cell mass where the random inactivation of the maternal (Xm) or the paternal X occurs (Okamoto et al., 2004; Mak et al., 2004). Afterwards, the inactive state is extremely stable and inherited during cell cycle. Reversion of this random X-inactivation only occurs in primordial germ cells. In the laboratory, we work on cell lines derived from the three lineages of the blastocyst (ES, TS and XEN cells (Kunath et al. 2005)) as well as directly on developing embryos.
We ask: what are the factors or epigenetic marks ensuring a stable inactive state; whether these marks are uniformly distributed along the inactive X-chromosome and what ensures X chromosome plasticity in pre-implantation embryos? To do this, we are bringing together a cell biology analysis of X chromosome nuclear organisation using 3D fluorescence in situ hybridisation (FISH) and a biochemical study of chromatin structure on high-resolution microarrays (ChIP on chip + ChIPSeq).
Figure 2. Nuclear movements and association with nuclear compartments.
Gene organisation within the nucleus is not random. Active transcription units tend to loop out from their chromosome territory (left panel) to be recruited to or excluded from constitutive nuclear compartments (examples of nuclear compartments on middle panel; example of gene recruitment to transcription factories on the right panel). These nuclear movements are thought to exert a crucial function in epigenetic regulation of gene expression.
We are also interested in the relationship between the initial lineage commitments and changes in X-inactivation states, i.e. how the different forms of X-inactivation (imprinted, random and Xp reactivation) are established at the morula-blastocyst transition; what are the molecular triggers ensuring these changes? For these studies we make use of mouse models available in the laboratory and of genetically engineered cell lines where lineage commitment is controlled.
Figure 3. Relationship between lineage commitment and X-inactivation.
The left panel shows that X-inactivation (identified by the presence of a Xist domain) is induced only in cells which have lost the pluripotency marker Nanog (Navarro et al., 2008). Trophectoderm cells can be obtained by inducible overexpression of Cdx2 in ES cells (Niwa et al., 2005) (middle panel). In order to monitor X-inactivation in vivo in each cell layer, we use an Xp-linked GFP transgene (Hadjantonakis et al., 1998).
 

Dissecting functional elements within the Xic through targeted mutagenesis.

Philippe Clerc et al.


Background
We have shown critical regulatory sequences to be present within a 65 kb genomic span lying 3’ to Xist . This region is involved in the choice of which X chromosome to inactivate in each female cell (Clerc and Avner, 1998).

This panel shows DAPI-stained nuclei of differentiated ES cells carrying a 65 kb cre/loxP deletion in one of their two X-chromosomes. DNA-RNA FISH for Xist (green) and for the deleted region (red) detects the accumulation of Xist RNA on the deleted X-chromosome. Surprisingly in these cells, the un-mutated X, as localised by the red signal, is never coated by Xist RNA.

A site-specific cre/loxP re-insertion strategy has been used to identify functional elements within the 65 kb genomic span.

Using this approach in ES cells, we have shown that Tsix, a transcription unit antisense to Xist, is necessary but not sufficient for normal random choice (Morey et al., 2001).

On going analysis :
Targeting the 65 kb region for deletion in both XO (Morey et al 2001) and XY ES cells (Morey et al 2004) showed that this region mediates the counting process. Unlike wild-type cells, deleted male and XO differentiated ES cells initiate X-chromosome inactivation.
The targeting of smaller deletions and site-specific cre/loxP transgenes to within this region will allow the identification of the counting element.

Exploring the Xce

Philip Avner et al.


The X-controlling element (Xce) was originally defined by Cattenach in the 1960's as a genetic element influencing which of the two X chromosomes in the female cell would be inactivated. Early work showed that Xce mapped between the Tabby and Mottled (Atp7a) coat colour markers. In collaboration with Bruce Cattenach we have fine mapped Xce to within the Xic region lying centromeric to Xist (unpublished work). Studies on over 70 recombinant animals has enabled a candidate region of small size to be defined and candidate genes to be characterised. The candidate genes are being followed up by a multi-disciplinary approach for their role in the X-inactivation process. We hope that the end of the Xce enigma is nearing.

Post-transcriptional regulation of Xist and of the X-inactivation process

Philip Avner et al.


Previous work in the laboratory (Ciaudo et al., 2006) implicated genes in the nonsense mediated decay (NMD) pathway in the control of Xist regulation and X-inactivation. Whilst the precise mechanism by which this occurs remains obscure, the idea that an RNA processing mechanism would be involved in regulating Xist is inheritently appealing. Current experiments are aimed at establishing the mechanisms involved in this regulation and the eventual interactions between these mechanisms and the recently described regulation of Xist by pluripotency factor binding to Xist intron 1.

Controlling Xist expression at the transcriptional level

Philip Avner et al.

The laboratory has contributed largely to our knowledge of how the Tsix antisense and more recently pluripotency factors such as Oct 3/4 and Nanog contribute to controlling Xist transcription in undifferentiated and differentiating ES cells. Current work seeks to further clarify the mechanisms involved in the action of pluripotency factors in the context of other mechansims involved in X inactivation initiation such as counting and choice.