Women-only nights can be fun, but other events that occur exclusively to women are not so great. Why, for example, should only women who inherit a mutation in the BRCA1 tumour suppressor be prone to breast cancer? Shridar Ganesan et al. have shown, in the November 1 issue of Cell, that BRCA1 might be involved in X inactivation. Perhaps this could explain this female-specific phenomenon.

BRCA1 was known to localize to the unpaired X chromosome in pachytene spermatocytes, and Ganesan et al. confirmed this by showing that BRCA1 colocalized with a component of the XY body. The XY body shows similarities to the inactive X (Xi) chromosome in female somatic cells — both are heterochromatic, silenced and are coated with the non-coding XIST RNA — so does Xi also localize BRCA1?

Immunofluorescence of BRCA1 and fluorescent in situ hybridization (FISH) of XIST, carried out on female human cell lines, revealed that BRCA1 and XIST could colocalize to a nuclear structure — FISH of an X-chromosome probe confirmed that this was one of the X chromosomes. Chromatin immunoprecipitation analysis — using antibodies to BRCA1 or its binding partner BARD1 — followed by reverse-transcriptase PCR (RT-PCR) of XIST confirmed this interaction.

The next step was to investigate what happened in BRCA1-deficient cells. Frozen sections from sporadic breast and ovarian cancers were examined, and although the majority had nuclear BRCA1 and focal XIST (as opposed to diffuse) staining, those from BRCA1-deficient women did not. The HCC1937 human breast cancer cell line — which contains a germline mutation in one BRCA1 allele and has lost the wild-type allele — also lacked focal XIST staining. This could be restored by ectopic expression of wild-type BRCA1, but not of cancer-associated BRCA1 mutants. Similarly, RNAi of BRCA1 in wild-type cells decreased focal staining of XIST.

So how does BRCA1 regulate XIST — through its localization, synthesis or stability? The authors carried out RT-PCR of XIST in HCC1937 cells that were transfected with either a vector control or with wild-type BRCA1, to distinguish between these alternatives. The levels of XIST RNA were equivalent in both transfected lines, indicating that BRCA1 can influence XIST localization, but not its synthesis or stability.

As XIST is required for X-chromosome inactivation, Ganesan et al. investigated whether loss of BRCA1 influences the pattern of histone H3 methylation on lysine 9 (H3mK9), which is associated with transcriptional silencing. In female cells, there is a large amount of H3mK9 staining on Xi, but this is absent from HCC1937 cells. Similarly, H3mK9 immunofluorescence analysis on frozen sections of sporadic and BRCA1-deficient breast cancers indicates that BRCA1 is required for focal staining of H3mK9, and hence gene silencing.

But is loss of BRCA1 expression sufficient to reactivate previously silenced genes? This was tested in a female mouse cell line in which one X chromosome carried a non-functional copy of Xist and the other, inactivated, X chromosome carried a silenced copy that was tagged with GFP. RNAi of Brca1 resulted in the reactivation of Xist–GFP in a subset of these cells.

The loss of BRCA1 in women might therefore reactivate genes that are normally silent on the Xi. The upregulation of a set of X-chromosomal genes in BRCA1-deficient ovarian cancers lends support to the importance of this phenomenon in promoting tumorigenesis, but the establishment of a firm link remains a future goal.