1.2. Rett Syndrome
Rett syndrome (RTT; OMIM 312750) is an X-linked dominant postnatal progressive
neurodevelopmental disorder and the second most common cause of mental retardation
affecting females (Rett, 1966; Hagberg et al., 1983).
1.2.1. Historical Background
In 1965, Dr. Andreas Rett, an Austrian physician, investigated two disabled girls
having abnormal hand movements like winging, washing and clapping. One year later, he
reported 22 girls with these abnormal clinical features (Rett, 1966). However, the report
would not be recognized in the medical community for 17 years. In 1983, Dr. Bengt
Hagberg, a Swedish neurologist, and his colleagues reported the clinical description of 35
cases (Hagberg et al., 1983). He recognized that his patients showed overlapping
phenotype with Dr. Rett’s and called the disease as Rett Syndrome to honor the first
description by Dr. Rett.
1.2.2. Clinical and Neuropathological Charecteristics
After normal development up to the age of 6 to 18 months, RTT patients show a
regression of motor and mental abilities. The clinical manifestations in the classical form
of RTT are characterized by cognitive deterioration with autistic features, loss of acquired
skills such as language and hand usage, stereotypical hand wringing movements, and gait
ataxia (Weaving et al., 2005). Behavioral abnormalities include teeth grinding, night
laughing or crying, screaming fits, low mood, and anxiety episodes elicited by distressful
external events (Mount et al., 2001). Patients suffer generalized rigidity, dystonia, and
worsening of scoliosis. Autistic features include expressionless face, hypersensitivity to
sound, lack of eye-to-eye contact, indifference to the surrounding environment, and
unresponsiveness to social cues (Segawa and Nomura, 2005). Most girls with RTT suffer
7
breathing anomalies, including breath-holding, aerophagia, forced expulsion of air and
saliva, apnea, and hyperventilation.
Patients with RTT develop postnatal microcephaly. The major morphological
abnormalities detected in the central nervous system (CNS) are an overall decrease in the
size of the brain and of individual neurons. Autopsy studies revealed a 12–34 per cent
reduction in brain weight and volume in patients with RTT, the effect most pronounced in
the prefrontal, posterior frontal, and anterior temporal regions (Armstrong et al., 2005).
The RTT brain shows no obvious degeneration, atrophy, or inflammation, and there are no
signs of gliosis or neuronal migration defects (Jellinger et al., 1988; Reiss et al., 1993).
These observations indicate that RTT is a disorder of postnatal neurodevelopment rather
than a neurodegenerative process. In addition, dendritic spines of the RTT frontal cortex
are sparse and short, with no other apparent abnormalities (Belichenko et al., 1994).
Although neuronal size is reduced in the cortex, thalamus, basal ganglia, amygdala, and
hippocampus, there is an increase in neuronal cell packing in the hippocampus (Kaufmann
and Moser, 2000).
Neurophysiological studies of mouse models and patients with RTT revealed that
both the CNS and the autonomous nervous system contribute to the pathophysiology of the
disease. Altered somatosensory evoked potentials and abnormal electroencephalogram
(EEG) findings of focal, multifocal, and generalized epileptiform discharges and the
occurrence of rhythmic slow theta activity, all suggest altered cortical excitability in the
RTT brain. Electrocardiographic recordings (ECG) demonstrate long corrected QT
intervals and suggest perturbation of the autonomic nervous system (Glaze et al., 2005).
Mouse models reveal abnormalities in long-term potentiation (LTP) and impaired synaptic
plasticity. LTP is reduced in Mecp2
_/Y
and Mecp2
308/Y
cortical slices whereas LTP is
enhanced in hippocampal slices of mouse with MeCP2 overexpression (MECP2
Tg
) (Asaka
et al
., 2006; Moretti et al., 2006; Collins et al., 2004). In addition, reduced spontaneous
activity in cortical slices of null mice was observed due to a decrease in the total excitatory
synaptic drive and an increase in the total inhibitory drive (Dani et al., 2005). Synaptic
outputs in glutamatergic neurons showed 50 per cent reduction and 100 per cent
enhancement in Mecp2
-/Y
and MECP2
Tg
, respectively (Chao et al., 2007).
8
Altogether, these findings indicate that MeCP2 is essential in modulating synaptic
function and plasticity, and that MeCP2 function is critical in regulating the number of
excitatory synapses during early postnatal development (Chahrour and Zoghbi, 2007).
1.2.3. Phenotypic Variability in RTT
Atypical variants of RTT are also commonly observed, and five distinct categories
have been delineated on the bases of clinical criteria: Infantile (early) seizure onset,
congenital forth, ‘forme fruste’, preserved speech variant (PSV), and late childhood
regression form (Hagberg et al., 2002). These variants range from milder forms with a later
age of onset to more severe manifestations. ‘Forme fruste’ comprises the most common
group (with 80 per cent) of atypical variants. These patients have surprisingly well
preserved, yet somewhat dyspraxic, hand function, as well as absence of the classic hand
wringing stereotypies (Hagberg, 2002). The PSV variant is characterized by the ability of
patients to speak a few words, although not necessarily in context. PSV patients have a
normal head size and are usually overweight and kyphotic (Zappella et al., 2001). Early
seizure onset type and congenital forth are the more severe variants of RTT. Early seizure
onset type is characterized by a lack of early normal period due to presence of seizures
whereas congenital forth patients lack the early period of normal development (Chahrour
and Zoghbi, 2007). A definite loss of acquired hand skill is not found in congenital forth
cases, instead an improvement with age in their most primitive early bilateral hand use is
seen (Hagberg et al., 2002). Classical and atypical RTT phenotypes vary in severity and
onset between different patients and in the same patient over time.
1.2.4. Genetic Basis of RTT
RTT has an incidence of 1/10,000 to 1/22,000 female live births (Percy, 2002).
However, since more than 99 per cent of RTT cases are sporadic, it was very hard to map
the disease locus by traditional linkage analysis. Using information from rare familial
cases, exclusion mapping identified the Xq28 candidate region, and subsequent screening
of candidate gene, methyl-CpG-binding protein 2 (MECP2; MIM# 300005), revealed
mutations in RTT patients (Amir et al., 1999). Several recent studies identified mutations
in the CDKL5 gene (OMIM 300203) encoding cyclin-dependent kinase like 5 in patients
9
with an atypical, early onset seizure variant of RTT (Weaving et al., 2005; Evans et al.,
2005; Scala et al., 2005). The disruption of NTNG1 gene, encoding the axon guiding
molecule Netrin G1, by a balanced chromosome translocation was described in one female
patient with atypical RTT and early-onset seizures (Borg et al., 2005). However, this might
be an isolated case because NTNG1 screening in a cohort of MECP2 and CDKL5 mutation-
negative patients with RTT failed to identify any pathogenic mutations in this gene (Archer
et al
., 2006).
RTT was initially thought to affect exclusively females and germline MECP2
mutations were considered to be lethal in males. Recently, several investigators have
reported MECP2 mutations in males with classic RTT, nonfatal nonprogressive
encephalopathy, nonspecific X-linked mental retardation (MRX), language disorder, or
schizophrenia (Orrico et al., 2000; Cohen et al., 2002; Kleeftra et al., 2004; Masuyama et
al
., 2005). Males with MECP2 mutations fall into three main categories: boys with Rett
syndrome; boys with severe encephalopathy and infantile death; and boys with less severe
neurological and/or psychiatric manifestations. Boys in the first category have a 47,XXY
karyotype or are somatic mosaic and carry the same MECP2 mutations that cause classic
Rett syndrome in girls. Males in the second group carry MECP2 mutations identical to
those found in females; these mutations are generally thought to disrupt DNA binding or
nuclear localization of the MECP2 protein. In the third group are boys with mutations that
are not found in girls with Rett syndrome, presumably because their effects are mild in
heterozygosity. Recent data indicate that increased MECP2 gene-dosage can disrupt
normal brain function. Interestingly, submicroscopic duplications in Xq28 region
encompassing the MECP2 gene were identified in a boy with severe mental retardation and
clinical features of Rett syndrome, several patients with severe mental retardation and
progressive spasticity, and male with non-specific X-linked mental retardation (Meins et
al
., 2005; van Esch et al., 2005; de Gaudio et al., 2006; Friez et al., 2006).
1.2.5. MECP2 Gene
The MECP2 gene is located on chromosome Xq28 and consists of four exons that
code for two different isoforms of the protein, due to alternative splicing of exon 2. Two
MeCP2 isoforms differ only in their N-terminus. The first identified isoform, MeCP2-e2
10
(MeCP2A, 486 amino acids) uses a translational start site within exon 2, whereas the new
isoform, MeCP2-e1 (MeCP2B, 498 amino acids) is derived from an mRNA in which exon
2 is excluded and starts from ATG located within exon 1. MeCP2-e1 isoform is more
abundant and contains 24 amino acids encoded by exon 1 and lacks the 9 amino acids
encoded by exon 2 (Dragich et al., 2007; Kriaucionis and Bird, 2004; Mnatzakanian et al.,
2004). In addition, MECP2 has a large, highly conserved 3’-untranslated region that
contains multiple polyadenylation sites, which can be alternatively used to generate four
different transcripts. Expression studies in mice showed that the longest transcript is the
most abundant in brain, with higher expression during embryonic development, followed
by postnatal decline, and subsequent increase in expression levels later in adult life (Pelka
et al
., 2005; Shahbazian et al., 2002b). Although the MeCP2 is almost ubiquitously
expressed, it is relatively more abundant in the brain, primarily in mature postmigratory
neurons (Jung et al., 2003). MeCP2 protein levels are low during embryogenesis and
increase progressively during the postnatal period of neuronal maturation (Balmer et al.,
2003; Cohen et al., 2003; Kishi and Macklis, 2004; Mullaney et al., 2004; Shahbazian et
al
., 2002b). Both MeCP2 isoforms are nuclear and colocalize with methylated
heterochromatic foci in mouse cells. Since MeCP2 is expressed in mature neurons and its
levels increase during postnatal development, it may play a role in modulating the activity
or plasticity of mature neurons. Consistent with this, MECP2 mutations do not seem to
affect the proliferation or differentiation of neuronal precursors.
MeCP2 is a member of the family of related proteins that bind specifically to
symmetrically methylated CpG dinucleotides via a conserved methyl binding domain
(MBD) (Bird, 2002). Besides a methyl binding domain (MBD, residues 78–162), the
protein includes a transcription repression domain (TRD, residues 207–310) involved in
transcriptional repression through recruitment of co-repressors and chromatin remodeling
complexes, two nuclear localization signals (NLS, residues173-193 and 255–271), and a
63 residue group II WW binding domain in C-terminal (from 325 to 388) (Figure 1.3). The
C-terminus facilitates MeCP2 binding to naked DNA and to the nucleosomal core, and it
also contains evolutionarily conserved poly-proline runs that can bind to group II WW
domain splicing factors (Buschdorf and Stratling, 2004).
11
Figure 1.3. The MECP2 gene (a) and its protein product (MeCP2A) with conserved
domains (b).
1.2.6. CDKL5 Gene
CDKL5 gene, previously known as serine/threonine kinase 9 (STK9), is located on
chromosome Xp22. Alterations in this gene were originally found to cause early-onset
epilepsy and infantile spasms with severe mental retardation (Grosso et al., 2007). The
observation that mutations in MECP2 and CDKL5 cause similar phenotypes in early onset
seizure variant of RTT suggested that these genes might participate in the same molecular
pathways. Mari et al. (2005) showed that MECP2 and CDKL5 have an overlapping
temporal and spatial expression profile during neuronal maturation and synaptogenesis and
that they physically interact. The interaction was shown to require a portion of the C
terminal domain of MeCP2, suggesting that mutations in this region might be involved in
RTT onset owing to loss of interaction between the two proteins. Furthermore, it was
shown that the kinase activity of CDKL5 can cause both autophosphorylation and MeCP2
phosphorylation, and this latter activity is eliminated in pathogenic CDKL5 mutants
(Bertani et al., 2006). Phosphorylation of MeCP2 has a crucial role in the regulation of its
target gene expression (Mari et al., 2005).
1.2.7. MECP2 Mutation Profile
Mutations in the MECP2 gene are associated with several disorders that include Rett
syndrome (RTT), Angelman syndrome like phenotype, autism, and even mild forms of
12
mental retardation (Amir et al., 1999; Couvert et a., 2001; Watson et al., 2001). MECP2
mutations can be identified in 70–90 per cent of classical sporadic RTT cases, however, in
only 29-45 per cent of atypical RTT and familial cases (Webb and Latif, 2001; Schanen et
al
., 2004, Fukuda et al., 2005).
Up to date more than 200 different mutations of MECP2 have been identified in
patients with classical and atypical RTT. Most mutations are de novo that arise in the
paternal germline and often involve a C to T transition at CpG dinucleotides (Trappe et al.,
2001; Wan et al., 1999). Among 200 different mutations of MECP2 eight missense and
nonsense mutations (p.R106W, p.R133C, p.T158M, p.R168X, p.R255X, p.R270X,
p.R294X, and p.R306C) are known to account for almost 70 per cent (RettBASE).
Among those mutations, the great majority (80 per cent) represent single nucleotide
changes, with the remainder small-scale deletions (17 per cent) or insertions (3 per cent).
Most missense mutations are tightly clustered at the methyl-CpG binding domain (MBD).
Deletion/insertion mutations leading to loss of the open reading frame occur throughout the
gene, but are clustered in the C-terminal coding region, which contains a poly-histidine
repeat (RettBASE).
With the use of quantitative experimental methods in recent years, MECP2 exon
deletions were identified in 2.9-14 per cent of cases with RTT (RettBASE). On the other
hand, duplication was reported in only one female patient with PSV variant of RTT who
was found to carry three copies of the MECP2 exon 4 (Ariani et al., 2004). Regardless of
the mechanism, mutations within the MECP2 lead to loss of function or a protein product
with diminished stability.
A wide spectrum of phenotypic variability is observed in patients with MECP2
mutations and considered with respect to the mutation type, location in the gene, and the
X-chromosome inactivation (XCI) pattern. Genotype–phenotype correlations in females
with Rett syndrome have yielded conflicting results. In general, female patients with
mutations in MECP2 that truncate the protein towards its C-terminal end (late-truncating
mutations) have a milder phenotype, and less typical of classical Rett syndrome when
compared to patients who have missense or N-terminal (early truncating) mutations
13
(Charman et al., 2005). Patients with mutations upstream of or within the TRD domain
show greater clinical severity (Jian et al., 2005). In addition, p.Arg133Cys mutation causes
an overall milder phenotype while the p.Arg270X mutation, which is predicted to result in
a truncated protein, is associated with increased mortality (Kerr et al., 2006; Jian et al.,
2005).
It has been suggested that genetic background and/or non-random X-chromosome
inactivation in the brain influences the biological consequences of mutations in MECP2. In
females, only one of the two X chromosomes is active in each cell and the choice of which
X chromosome is active is usually random, such that half of the cells have the maternal X
chromosome and the other half have the paternal X chromosome active. Therefore, a
female with a MECP2 mutation is typically mosaic, whereby half of her cells express the
wild-type MECP2 allele and the other half express the mutant MECP2 allele. Occasionally,
cells expressing the wild-type MECP2 allele divide faster or survive better than cells
expressing the mutant allele, which therefore results in a nonrandom pattern of XCI and
amelioration of the RTT neurological phenotypes. Depending on the extent of such
favorable skewing, some patients can be mildly affected or are even asymptomatic carriers
of MECP2 mutations (Weaving et al., 2005). The best examples for illustrating the
dramatic effects of XCI patterns in RTT are monozygotic twins who manifest very
different phenotypes (Dragich et al., 2000). In addition, skewed XCI patterns occur in
brain regions of female mice heterozygous for a mutant MECP2 allele, where phenotypic
severity correlates with the degree of skewing (Young and Zoghbi, 2004).
1.2.8. MeCP2 Function
1.2.8.1. Transcription Regulation and Chromatin Remodeling. Findings of extensive
research suggests that MeCP2 acts as a transmitter of epigenetic information by binding to
methylated CpG dinucleotides, recruiting complexes that include histone deacetylase and
methyltransferase, and leading to local transcriptional repression. The function of MeCP2
as a transcriptional repressor was first suggested based on in vitro experiments in which
MeCP2 specifically inhibited transcription from methylated promoters (Nan et al., 1997).
When MeCP2 binds to methylated CpG dinucleotides of target genes via its MBD, its TRD
recruits the corepressor Sin3A and histone deacetylases (HDACs) 1 and 2 (Jones et al.,
14
1998; Nan et al., 1998). The transcriptional repressor activity of MeCP2 involves
compaction of chromatin by promoting nucleosome clustering, either through recruitment
of HDAC and histone deacetylation or through direct interaction between its C-terminal
domain and chromatin (Figure 1.4) (Nikitina et al., 2007).
According to the dominant model of MeCP2 action, target genes are silenced by
MeCP2 binding to the promoter. However, combined ChIP–chip promoter and expression
profiling analysis reveals that 62.6 per cent of MeCP2-bound promoters (including BDNF)
are transcriptionally active (Yasui et al., 2007). These studies clearly demonstrate that
MeCP2 promoter occupancy does not correlate with transcriptional silencing of target
genes but rather functions as a modulator of gene expression depending on the physiologic
state of the organism. Metaphorically speaking, MeCP2 may be best thought of as the
dimmer that regulates the amount of light rather than the switch that turns the lamp on and
off (Chahrour and Zoghbi, 2007). Extensive studies of the binding of MeCP2 (or the MBD
alone) to DNA in vitro have revealed that the affinity for methylated DNA is not strong
and is only ~3-fold weaker for unmethylated DNA (Fraga et al., 2003).
Transcriptional profiling indicates that MeCP2 is not a general transcriptional
repressor in vivo but has a more subtle effect involving a subset of genes (Tudor et al.,
2002; Ballester et al., 2005). Also, the finding that MeCP2-induced repression is only
partially alleviated by inhibiting HDACs suggests that its activity is not restricted to
HDAC recruitment (Yu et al., 2000). Furthermore, there is evidence that MeCP2 is
responsible for the formation of large chromatin loops (Horike et al., 2005) (Figure 1.5).
Nikitina et al. (2007) has shown that MeCP2 dependent chromatin loop formation occurs
in two steps: a methylation-independent interaction between chromatin and the C terminus
that is required for the second, methylation-specific, interaction between DNA and the
MBD domain. The yellow and blue arrows in Figure 1.5 indicate MeCP2-interacting
sequences. When MeCP2 is present, it interacts with sequences that are near the imprinted
DLX5 and DLX6 genes and define the boundaries of an 11-kb chromatin loop. This leads
to an integration of DLX5 and DLX6 into a loop of silent, methylated chromatin, and
represses their expression (Figure 1.5a). In neurons that are deficient for MeCP2, the
chromatin in this region is structured into a distinct conformation that corresponds to active
chromatin loops, which are bordered by sequences (indicated by long purple and orange
15
arrows) that interact with chromatin factors. Therefore, in MeCP2- deficient neurons, the
expression of DLX5 and DLX6 is no longer repressed (Figure 1.5b) (Bienvenu and Chelly,
2006).
Figure 1.4. Mechanisms of methylation dependent (a) and independent (b) transcription
regulation and chromatin remodeling (Bienvenu and Chelly, 2006).
1.2.8.2. RNA Splicing. Recent studies suggest that the function of MeCP2 might be more
complex than previously anticipated. For instance, purified recombinant MeCP2 was
shown to have a high-affinity RNA binding activity that is mutually exclusive to its
methyl-CpG-binding properties and does not require the methyl-CpG-binding domain
(Jeffery et al., 2004).
16
Figure 1.5. Regulation of imprinted regions through formation of a silent chromatin loop.
(a) transcriptionally inactive and (b) active conformation (Bienvenu and Chelly, 2006).
Interestingly, although the biological significance of a MeCP2–RNA complex
remains to be elucidated, recent data indicated that MeCP2 interacts with the RNA-binding
Y box-binding protein 1 (YB1) and regulates splicing of reporter minigenes (Figure 1.6)
(Young et al., 2005). Importantly, aberrant RNA-splicing patterns of several genes
including Dlx5 were identified in Mecp2 null mice (Young et al., 2005). The finding that
MeCP2 regulates transcription and splicing of some of its targets suggests the existence of
multiple layers of epigenetic regulation (Moretti and Zoghbi, 2006).
Figure 1.6. Regulation of alternative splicing by MeCP2; a) RNA splicing in the presence
of MECP2 and b) aberrant splicing in the absence of MECP2 (Bienvenu and Chelly, 2006).
|