Hippocampal CA1/CA2 neurons from a normal (left) and a Mecp2-null mouse, showing its smaller neurons. Credit: Courtesy of R. Chen and R. Jaenisch, Whitehead Institute, USA. Reproduced with permission from Nature Genetics © Macmillan Magazines Ltd (2001).

Rett syndrome — an inherited, X-linked neurological disorder — is characterized in females by 7–18 months of seemingly normal postnatal development, followed by a rapid neurological deterioration that leads to dementia, autism, loss of speech and voluntary movement, and 'acquired' microcephaly. Rett is lethal to males but females can survive these crises and live for several decades.

In 1999, female Rett patients were found to be heterozygous for mutations in Mecp2 , which encodes a methyl-CpG binding protein that acts as a transcriptional repressor in chromatin remodelling complexes. Mecp2 was then believed to be essential for embryogenesis as attempts to produce mice from Mecp2-null ES cells had failed. This raised a key question: Is Rett a prenatal developmental disorder, which manifests postnatally, or is it due to the malfunctioning of postnatal neurons? Now two papers shed light on this question with the conditional inactivation of Mecp2.

Both groups used Cre-loxP technology to generate mice that lack Mecp2 by crossing a 'floxed' Mecp2 allele onto mice that ubiquitously express Cre. The resulting animals produced Mecp2-null offspring of both sexes that were viable and seemingly normal, proving that this gene is not required for embryogenesis. However, by five weeks of age, null mice developed Rett-like movement, breathing and behavioural abnormalities and died around five weeks later. As in female Rett patients, Guy et al. found that neurological deterioration in Mecp2+/- took six months to develop, reproducing a key feature of the disease. But the brains of their null mutant mice appeared grossly normal at autopsy, although Chen et al. report that their null mutants had brains slightly smaller than normal.

This phenotype is due to the loss of Mecp2 in the brain alone, because mice that lack Mecp2 mainly in the CNS (generated by expressing Cre from the promoter of the CNS-expressed nestin gene) developed with an identical phenotype to that of the null Mecp2 mice. Neurons in the brains of the null and nestin-Cre Mecp2 mutants had cell bodies and nuclei that were smaller and more closely packed together in certain brain regions than wild-type mice (see picture) — changes also seen in Rett patients. These findings indicate that Mecp2 might help maintain neuronal function. Indeed, the phenotype of mice in which Mecp2 was removed only from postnatal neurons was similar to that of Mecp2-null mice, although less severe and later in onset. Importantly, these findings show that loss of Mecp2 in postnatal neurons, and not in glial cells, produces a Rett phenotype in a cell-autonomous manner.

These findings augur well for developing future therapies for Rett syndrome as the neurological disturbances that occur might not be due to abnormal brain development but to defective neuronal function caused by MECP2 loss. If it is a prolonged deficiency of MECP2 that triggers the neurological crisis then therapeutic strategies could target this deficiency postnatally. However, questions remain. How does loss of a ubiquitously expressed transcriptional repressor bring about changes in gene expression that result in a CNS-specific phenotype? And curiously, why do the Mecp2-null mice become obese? No doubt, these mice will provide many future questions and answers on Mecp2 and its role in Rett syndrome.