DNA methylation and the control of gene expression on the human X chromosome

In spite of the rather intense interest in the biology of X inactivation, little experimental data have been obtained that provide insight into the molecular basis for the initiation or the maintenance of X inactivation. Indeed, although the preponderance of data would indicate that supernumerary X chromosomes are inactivated, some investigators still consider it possible that the differentiative event is activation of a single X chromosome with the other X chromosomes remaining 'passively' inactive. Furthermore, it is clear that active X-encoded genes are not constitutively expressed in all cell types, and it is therefore probable that further levels of control of gene expression are superimposed on inactivation. In addition, it is not unlikely that the initiation of X inactivation will turn out to have a mechanism different from the process responsible for the maintenance of an inactivation pattern once it is established. At the time of inactivation, a chromosomal event clearly appears to be implicated. However, as discussed, individual segments and, perhaps, even individual genes can be independently 'reactivated' in vitro. Finally, it is possible that there may be several types of X inactivation, as perhaps is exemplified by the preferential and partial inactivation of the parental X chromosome in marsupials as opposed to the random and complete inactivation of the X chromosome seen in eutherian mammals. Preferential paternal X inactivation may also exist in mouse extraembryonic tissues, and a reversible form of X inactivation seems to occur in female primordial germ cells. Basically, three general types of models have been put forth in attempts to encompass and explain numerous biological observations regarding mammalian X-chromosome inactivation. These include specific DNA-protein interaction, physical rearrangement of the inactive X genes, and DNA modification.