Introduction - Rett Syndrome
Clinical signs
Rett syndrome was first described by Andreas Rett in 1966 (Rett, 1966). It accounts for 2 to 3% of all cases of severe mental retardation and 10% of such cases in women. Despite these important numbers, the pathological mechanisms are still unknown. Clinically, the affected girls have a normal in utero development followed by a normal postnatal period extending up to 18 months of age. Their development then slows down and stops. A rapid regression period then occurs with the loss of acquired skills. The patients have severe mental retardation and show a sudden arrest of brain development (the girls will have acquired microcephaly). Additional problems are present in these patientss (Rett Syndrome Diagnostic Criteria Working Group, 1988), among which : scoliosis, spasticity, breathing anomalies (apnea and hyperventilation episodes), epilepsy, abnormal blood circulation.
Genetics
A genetic origin to the disease was suspected early because of the three following reasons (Zoghbi, 1988) :
- only girls are affected
- rare familial cases exist with maternal transmission
- the phenotype of monozygotic twins is concordant
The possibility that this disease could be X-linked dominant with male letality was further reinforced by the fact that several male patients inside Rett syndrome families were affected by a severe encephalopathy which lead to their death during the first year of life (Schanen et al., 1998). Exclusion mapping was performed using less than a dozen of familial cases (with at least two affected child in the same family). It showed that the only region that could contain a causative gene defect was located at the tip of the long arm of the human X chromosome (Xq28) (Schanen et al., 1997 ; Sirianni et al., 1998).
Human X chromosome.
Identification of the MeCP2 gene
A systematic search for mutations in the genes of this region was undertaken and lead, during the fall of 1999,
to identify mutations in 30% of Rett syndrome patients in the MeCP2 gene (Methyl-CpG binding protein 2)
(Amir et al., 1999). Since then, a mutation in this gene was found in more than 95% of the patients. More than 500 mutations
are currently listed in the literature.
The MeCP2 gene was first identified in mice in 1992 by the group of Adrian Bird
(Lewis et al., 1992). The human gene was cloned in 1996
(D’Esposito et al., 1996). It spans 76 kilobases of genomic DNA between the genes coding for the interleukine receptor
associated kinase (IRAK) and a color vision gene (RCP). It is composed of 4 exons which are transcribed from telomere
to centromere. The open reading frame has a size of 1461 nucleotides and encodes a 486 amino acids MeCP2 protein.
It is ubiquitously expressed
(Reichwald et al., 2000) and it is subject to X-chromosome inactivation.
In early 2005, an alternative form of the MeCP2 protein was described. This transcript initiates traduction in exon 1 and does not
contain exon 2. The resulting protein has a different size (498 amino acids instead of 486) and has a different N-terminus
(21 amino acids)
(Mnatzakanian et al., 2005).
Structure of the MeCP2 gene, shown here with its 4 exons. The two coding transcripts (with and without exon 2) are labelled A and B, the arrows show the position of the traduction initiation and termination codons. The three polyadenylation sites are labelled pA1, pA2 and pA3. This gene is thus potentially producing 6 different transcripts.
In mice, the Mecp2 gene is expressed both in the embryo and in the adult. However, the expression level is low during the early phases of development (Meehan et al., 1992). In humans, three different transcripts are present (1.8, 7.5 and 10 kb) and they are generated by the use of three different polyadenylation signals in the 3’ non translated region. The functional significance of these three forms is currently not known. The functional studies that were conducted in mice have lead to important findings concerning the function of this protein.
Fonction of the MeCP2 protein
MeCP2 is a member of the family of proteins able to bind to methylated DNA on CpG dinucleotides. These dinucleotides are largely distributed in the mammalian genomes but they are enriched in regions which are transcriptionnally silent (heterochromatin) and in the promotor region of many genes. Approximately 60% of cytosines located in CpG pairs are methylated. This constitutes an essential mechanism to repress transcription both in the heterochromatin and in regions which are transiently repressed. Methylation itself is not suffucient to repress the expression of a gene. It was demonstrated that the onset of repression is possible because additional proteins are recruited at these methylated sites. For a review, see Ulrey et al., 2004.
The MeCP2 protein contains two functional domains : a methyl binding domains (MBD) which binds methylated CpG dinucleotides, and a transcriptional repression domain (TRD) ( Nan et al., 1993 ; Nan et al., 1998). Co-immunoprecipitation experiments showed that MeCP2 is associated, in mice, with several other proteins such as Sin3A (a transcriptional co-repressor) and the histone deacetylases HDAC1 and HDAC2. MeCP2 would promote transcriptional repression through its binding to CpG dinucleotides and its subsequent recruitment of histone deacetylases that would modify the chromatin structure to make it inactive (this mechanism is presented below) :
Putative, "historical" model for the function of MeCP2. The MeCP2 protein binds methylated cytosine residues (mC) using its MBD. It recruits Sin3A using its TRD which, in turns interacts with the histone deacetylases (HDAC). Deacetylation of the histone tails compacts the chromatin and silences transcription. (based on Van den Veyver et Zoghbi, 2000).
A search for targets of the MeCP2 protein did not allow to detect massive deregulation of gene transcription when the MeCP2 gene product is absent. This absence of "global" modification of transcriptional activity in the affected cells is partly contradictory with the proposed role of this protein. A recent study showed that the MeCP2 protein is able to regulate the promotor III of the brain derived neurotrophic factor (BDNF) gene, a gene responsible for the production of a protein essential for the survival and growth of neurons (Martinowich et al., 2003 ; Chen et al., 2003). It is thus possible that the MeCP2 protein is not a global transcriptional regulator. The protein could have a reduced number of targets in cells and hence have a more subtle role than what was initially thought. More recently (Horike et al., 2005) a slightly different mechanism was proposed. Indeed, the MeCP2 protein could repress transcription via the formation of inactive DNA loops rather than a via direct repression like the mechanism proposed in the figure above. The role of the MeCP2 protein would thus be less simple than initially anticipated (binding upstream elements and silencing downstream genes). The protein could be able to recognize specific DNA sequences and would promote the creation of loops between these regions. The consequence would be the silencing of the genes contained inside the loops. One interesting point (in addition to the discovery of this alternative function) is that one such loop contains the Dlx5 gene. Dlx5 is important for the neurotransmitter GABA to be synthetized. A first draft is thus maybe appearing linking MeCP2, inactive DNA loops, Dlx5 and neurones (via GABA).
Alternative mechanism for MeCP2 mediated transcriptional repression. The protein could bind several DNA sequences and promote the formation of inactive DNA loops between these elements. The genes inside the loop are transcriptionally silent (based on Horike et al., 2005).
These two mechanisms are not necessarily mutually exclusive. A recent review on the subject can be found here (Chadwick et Wade, 2007).