Rett syndrome is a brain disorder caused by disrupted forms of the protein MECP2, but how MECP2 loss affects the brain is unknown. A mouse study now implicates key domains of the protein and offers therapeutic insights. See Letter p.398
Rett syndrome is a rare and devastating condition that results in intellectual and physical disability, and mainly affects females. Nearly all diagnosed cases are associated with mutations in the MECP2 gene that cause its encoded protein to become non-functional or to be present in reduced amounts1. However, because the MECP2 protein interacts with many other proteins and has both broad and specific roles in the regulation of gene expression, determining how MECP2 loss causes Rett syndrome has been difficult. On page 398, Tillotson, et al.2 report that expression of a version of mouse MeCP2 that contains just two key domains is associated with reduced symptoms in a mouse model of Rett syndrome, a finding that has therapeutic implications.
Rett syndrome is usually diagnosed when atypical movement and cognitive symptoms appear in infants at 6–18 months of age. It was initially assumed that the neurological symptoms of this condition reflect abnormal brain development that would probably be challenging to reverse. However, some experiments using animal models of the disease suggest that effective therapeutic intervention might be possible.
Mice genetically engineered to carry disease-associated mutations in the Mecp2 gene3 provide models for studying Rett syndrome; the animals display key characteristics of the condition, including movement and breathing abnormalities and premature death in males. When genetic engineering was used to induce MeCP2 expression in adult mice that had lacked the protein, some of the most severe effects of the syndrome, including the incidence of movement abnormalities and premature death, were reduced4. Another study5 confirmed that MeCP2 expression is required for normal brain function, because deleting Mecp2 in adult mice induced behavioural problems and premature death. Such studies raised hopes that, even after symptoms have appeared, therapies to restore MECP2 function might improve the lives of people with Rett syndrome6. However, restoring MECP2 expression in patients poses technical challenges, and the uncertainty about the specific way in which the protein prevents the condition is another hurdle to overcome.
MeCP2 contains several different domains, each with distinct roles. For example, the methyl-DNA binding domain (MBD) binds cytosine DNA bases that have a methyl group attached. Methylation of DNA that is in complex with histone proteins in a structure called chromatin is often associated with gene-expression changes, and it is presumed that MeCP2 can aid this regulatory process3. Other domains in MeCP2 are evolutionarily conserved, and mediate many other types of molecular interaction; for example, the protein's AT-hook domains can directly contact DNA. MeCP2 can also bind proteins that regulate gene expression through a range of mechanisms, including chromatin modification and RNA regulation3.
Although these observations indicate that MeCP2 functions as a scaffold to coordinate many processes of gene regulation, most missense mutations associated with Rett syndrome (those in which one amino acid is different from the wild-type version) occur either in the MBD or in an NCoR/SMRT interaction domain (NID). This domain interacts with a complex containing the proteins NCoR and SMRT, which are associated with inhibition of gene expression7. Tillotson and colleagues tested the hypothesis that the MBD and NID are the key MeCP2 domains relevant to Rett syndrome by genetically engineering mouse strains to express shortened versions of the protein.
One engineered strain contained MeCP2 missing its amino terminus (ΔN), another lacked both the amino and carboxy termini (ΔNC), and the smallest version (ΔNIC) contained just the MBD and NID (Fig. 1). Male mice that completely lack Mecp2 have a median survival of nine weeks and display severe movement and behavioural abnormalities. However, in the authors' experiments, the movement and behaviour of male mice that carried just ΔN or ΔNC were indistinguishable from those of their wild-type siblings. The strain carrying ΔNIC had a normal rate of survival, although these mice showed mild neurological deficits and reduced motor coordination compared with their wild-type siblings. Moreover, the authors demonstrated that, if ΔNIC was expressed in adult male mice lacking Mecp2, symptom progression was reduced and their survival rate was normal.
These results have implications for gene-therapy approaches to treat loss-of-function mutations in neurodevelopmental disorders. One challenge with such therapy is finding a suitable vector in which to transfer the required gene sequences. Self-complementary adeno-associated viruses (scAAVs) can be used for in vivo gene delivery, but these can package only small amounts of DNA, limiting their effectiveness to genes that encode small proteins or essential domains of larger genes.
To test this gene-therapy approach in their animal models, Tillotson and colleagues injected newborn male mice lacking Mecp2 with an scAAV that contained ΔNIC. This treatment increased the animals' rate of survival and reduced their Rett-syndrome-like symptoms compared with control mice that did not receive this version of MeCP2. Such an approach for investigating domains essential for protein function might also be useful in studies involving large, multi-domain scaffolding proteins — such as those of the SHANK protein family, which has been implicated in autism spectrum disorders8.
Tillotson and colleagues' work raises many questions, however. For example, is the recruitment of NID-binding proteins to methylated DNA required to support brain function? Perhaps methylation-dependent gene repression by the NCoR–SMRT complex is required. Or the NID might be needed to regulate the derepression of genes in a temporally precise manner, given that the neuronal-activity-dependent addition of a phosphate group to an amino acid in the NID disrupts its interaction with NCoR9. Determining these sorts of details about MeCP2 function will be crucial, because different drugs would be required to treat different modes of action of MeCP2.
Another complication in investigating Rett syndrome arises from the fact that the MECP2 gene is carried on the X chromosome. Males have only one X chromosome and show more severe symptoms for any given MECP2 mutation. So although the male mice used by Tillotson and colleagues provide a sensitive way of testing MeCP2 function, they have characteristics that are less pronounced in females. However, using female animals to investigate MeCP2 function can be tricky. Females have two X chromosomes, and, in any given cell, genes are expressed from only one of the chromosomes through a random process of inactivation. If a female has one mutant copy of the gene, this creates a mosaic of cells containing either wild-type or mutant MeCP2, so the severity of the symptoms of Rett syndrome tracks with both the random cellular pattern of X chromosome expression and the extent of MeCP2 impairment.
In future, it will be necessary to investigate whether the reversal of Rett-syndrome symptoms mediated by ΔNIC also occurs in female mice. Another aspect worth further study is whether symptom characteristics in females reflect compensatory changes in the cells that express normal MeCP2, in addition to the deficits in cells expressing mutant MeCP2. Indeed, in female mice10, messenger RNA profiles differ in neurons expressing either normal or mutant protein, compared with such profiles of neurons from mice carrying only a normal version of MeCP2. Given the protein's multifunctional nature, another interesting avenue for exploration is whether distinct domains are required for specific roles, such as cognitive or motor function. Studies that extend the genetic and viral approaches taken by Tillotson and colleagues will be essential to determine in more detail which functions of MECP2 offer the greatest promise for the development of treatments for Rett syndrome.Footnote 1
Krishnaraj, R., Ho, G. & Christodoulou, J. Hum. Mutat. 38, 922–931 (2017).
Tillotson, R. et al. Nature 550, 398–401 (2017).
Lyst, M. J. & Bird, A. Nature Rev. Genet. 16, 261–275 (2015).
Guy, J., Gan, J., Selfridge, J., Cobb, S. & Bird, A. Science 315, 1143–1147 (2007).
McGraw, C. M., Samaco, R. C. & Zoghbi, H. Y. Science, 333, 186 (2011).
Katz, D. M. et al. Trends Neurosci. 39, 100–113 (2016).
Lyst, M. J. et al. Nature Neurosci. 16, 898–902 (2013).
Monteiro, P. & Feng, G. Nature Rev. Neurosci. 18, 147–157 (2017).
Ebert, D. H. et al. Nature 499, 341–345 (2013).
Johnson, B. S. et al. Nature Med. http://dx.doi.org/10.1038/nm4406 (2017).