The XCI status of different

The XCI status of different hESCs was previously analyzed mainly by XIST papain inhibitor and was not conclusive (Hoffman et al., 2005; Adewumi et al., 2007; Shen et al., 2008; Silva et al., 2008; Dvash and Fan, 2009; Dvash et al., 2010). We now propose to characterize the status of X inactivation of multiple female cell lines by examining their transcriptional profile along the X chromosome and identifying the genomic regions with significantly higher levels of gene expression. This is possible by comparison of the gene expression profile of each female cell line to a reference baseline made up of a large number of male cell lines (Mayshar et al., 2010). We assume that regions within the female X chromosome, showing regional overexpression compared to the male dataset, represent active X chromosomes. In order to detect such regions, we implemented a piecewise constant fit (PCF) algorithm of the freely available comparative genomic hybridization (CGH) analysis software program, CGH-Explorer (Lingjaerde et al., 2005). Twenty-one female hESC lines were analyzed using this statistical test, enabling us to divide the XCI status of undifferentiated female hESCs into three categories: A, no XCI; B, full XCI; C, partial XCI (Fig. 1A). We defined cell lines with no XCI as those with overexpressing regions comprising over 85% of the chromosome. Cell lines with less than 15% overexpressing regions were defined as full XCI and lines with between 15 and 85% overexpressing regions were defined as partial XCI. Fig. 1B shows female hESC lines best representing each of the three XCI states compared to a representative male line in a moving average plot of the X chromosome. This classification is partially correlated with the average expression of XIST. Low expression of XIST is observed in all cell lines with no XCI, while in the cell lines with full or partial XCI XIST expression was either high or low (Fig. 1C). Thus, dissection of the cell lines according to XIST expression would create only two categories and would be misleading. Furthermore, we analyzed the active and inactive regions along the X chromosomes and generated an \"inactivation map\" for all the lines. As can be seen in Fig. 1A, the inactive regions of the chromosome in lines with partial XCI always include the regions adjacent to the chromosomal location of XIST. This finding is congruent with published models proposing that XIST spreads along the chromosome from its transcription site, where there is a high concentration of its RNA (Ng et al., 2007; Wutz et al., 2002). We therefore suggest that lines with partial XCI had undergone the first stages of inactivation with incomplete spreading of XIST. Interestingly, two subgroups of high and low expression levels of XIST are apparent in these lines. Since XIST is essential for the initiation of XCI but not for its maintenance (Silva et al., 2008; Csankovszki et al., 1999), we believe that cells with high levels of XIST are at the early stage of the process of X inactivation, whereas cells with low levels of XIST expression are at a later stage of the process, and probably involve epigenetic modifications preserving these regions on the X chromosome as inactive. The same explanation can be given for the same two subgroups of lines with full XCI. It still needs to be determined whether there are differences in the ability of these cell lines to initiate XCI on in vitro differentiation. High expression levels of pluripotent markers in all cell types (Fig. 1D) suggest that the differences in expression of genes on the X chromosome between the different cell lines did not result from differences in the differentiation status of the cells. All the cell lines tested show similar expression levels of their X chromosome pseudoautosomal genes (Fig. S1A), although their X-linked genes showed clear differences in their expression levels. As was recently suggested (Dvash et al., 2010), no correlation was found between the cell lines\' passage number or culture conditions and the XCI state (Table S1).