Portal de Periódicos da CAPES

Rett Syndrome and the Ongoing Legacy of Close Clinical Observation

2016; Cell Press; Volume: 167; Issue: 2 Linguagem: Inglês

10.1016/j.cell.2016.09.039

1097-4172

Huda Y. Zoghbi,

Family and Disability Support Research

This year marks the 50th anniversary of the publication of Andreas Rett's report on 22 girls who developed a peculiar and devastating neurological disorder that later came to bear his name. On this occasion, we reflect on the progress that has occurred in understanding Rett Syndrome, development of potential treatments, and the ramifications that Rett research has had on the fields of neurobiology and genetics. This year marks the 50th anniversary of the publication of Andreas Rett's report on 22 girls who developed a peculiar and devastating neurological disorder that later came to bear his name. On this occasion, we reflect on the progress that has occurred in understanding Rett Syndrome, development of potential treatments, and the ramifications that Rett research has had on the fields of neurobiology and genetics. As a young pediatrician, Andreas Rett chose to specialize in children with intellectual disabilities. Beginning in the mid-1950s, he closely followed a clinical population of over 6,000 children suffering various types of brain damage. He recognized that 22 of the children he observed—all girls—presented the same unusual symptoms with the same striking history. They had a normal birth. They reached early development milestones, some were perhaps even a bit precocious. Nine children did have some difficulty drinking in the early months, but for the most part the girls seemed healthy until at least 9 months of age, at which point they entered a period of functional decline that affected body, mind, and spirit. Their ability to stand was delayed, and walking even more so. Those children who did manage to walk had apraxia (they were unable to control the rhythm of their gait). All of the girls had an expressionless face with an empty gaze and little interest in interacting with their environment, including their parents. The minority who had learned a few mono- and di-syllabic words lost even this rudimentary vocabulary by the beginning of the third year of life. The most salient sign, however, was the stereotypical movements that occupied the girls' hands: from scratching, stroking, or kneading movements to clapping or flapping—always along the midline—the hands were in constant, but restricted, motion. Rett defined this new syndrome with nine obligatory features: it affects girls and causes flat affect, aphasia, hand stereotypies, hyperreflexia, spastic increase in tone, gait apraxia, a tendency to develop seizures, and profound cognitive impairment. Rett published his report in German in 1966 (Rett, 1966Rett A. Wien. Med. Wochenschr. 1966; 116: 723-726PubMed Google Scholar). A few isolated papers appeared over the next decade describing girls with overlapping features (reviewed in Hagberg, 1989Hagberg B.A. Pediatr. Neurol. 1989; 5: 75-83Abstract Full Text PDF PubMed Scopus (81) Google Scholar). But when the Swedish pediatric neurologist Bengt Hagberg met Andreas Rett at a medical conference, he realized the disorder that he and his coauthors had called "a progressive syndrome of autism, dementia, ataxia, and loss of purposeful hand use in girls," was the same disorder described by Rett (Figure 1). Hagberg christened the disorder Rett syndrome in 1983, in a paper describing 35 patients from France, Portugal, and Sweden (Hagberg et al., 1983Hagberg B. Aicardi J. Dias K. Ramos O. Ann. Neurol. 1983; 14: 471-479Crossref PubMed Scopus (1164) Google Scholar). Because it was published in the Annals of Neurology, awareness of the syndrome increased rapidly. If not for that fateful conference, it might have taken a very long time to realize these different small populations of girls suffered from the same disease, for Rett syndrome is very tricky to diagnose. If one encounters an infant who has just developed social withdrawal and hand stereotypies, one will think of autism. If, on the other hand, one encounters a girl at the stage when seizures and ataxia are prominent, one might think of an early-onset ataxia syndrome. A clinical diagnosis of Rett syndrome is difficult before 4 years of age, and this is not even taking into account the phenotypic variability of the disease (see below). My own relationship with Rett is further testament to serendipity. I was a resident in child neurology in fall of 1983 when I encountered my first patient with hand stereotypies and gait apraxia and loss of language, social, and cognitive function. She had been referred by Merlene McAlevy, a pediatrician who had seen Hagberg's paper and suspected the girl had Rett. As I read through the paper, it was clear that our patient matched every symptom described. The following week, I encountered a 12-year-old girl who had been diagnosed with cerebral palsy and recognized that she, too, had Rett. I began searching for similar cases and soon found six girls with Rett syndrome. My colleagues and I began officially recruiting and studying Rett patients in the U.S. International symposia on Rett syndrome were held in Vienna in 1984 and in Baltimore in 1985. Andreas Rett and Bengt Hagberg and several pediatric neurologists, neuropathologists, and neurophysiologists were in attendance, along with the newly diagnosed patients and their families. The meetings were both heart wrenching and fascinating. It was painful to see dozens of girls in one room wringing their hands, holding their breath or hyperventilating, grinding their teeth, and walking with a stiff apraxic gait; and it was even more painful to watch so many experts examining them with minute care yet having nothing to offer them or their families. I recall, however, the gentleness and respect with which Rett and Hagberg approached each of the girls and felt that such dedication and concern would ultimately lead to answers. After the careful evaluations, we would have long discussions about the potential causes and how to pursue biomarker studies. Early on, some neurologists were skeptical that Rett syndrome was a distinct entity rather than, for example, a type of cerebral palsy, but the consistently peculiar clinical presentation soon silenced the skeptics, and more and more patients were identified in various countries and among different populations. The prevalence of Rett syndrome is currently estimated at one in 10,000 females, making it one of the most common causes of intellectual and developmental disabilities in females. The chief obstacle to reliably diagnosing Rett syndrome, and thus being certain we were studying the same entity, was the absence of a biological marker. Although the disorder affected almost every aspect of nervous system function, there were no consistent changes that could be used as a diagnostic tool or as a point of entry to pathogenesis studies. Nevertheless, we quickly learned what Rett syndrome was not: it was not a urea cycle disorder, not a post-infectious or autoimmune disorder, not a metabolic disorder, not a degenerative disorder, and not a storage disorder. The absence of a useful neurochemical or biochemical alteration is not unique to Rett syndrome but actually rather common among intellectual disabilities and autism spectrum disorders. Even post-mortem studies were not terribly informative, although they confirmed the acquired microcephaly first noted by Rett. The brain weight tends to be ∼20% less than normal, with the cerebral hemispheres accounting for most of this reduction, which comes not from neuronal loss but from decreased dendritic arborization and more densely packed neurons. The size of the cortical minicolumns is also reduced in some areas, and the neurons of the substantia nigra pars compacta, although of normal number, have little to no neuromelanin granules (Armstrong, 2005Armstrong D.D. J. Child Neurol. 2005; 20: 747-753Crossref PubMed Scopus (176) Google Scholar). Electroencephalographic (EEG) studies, however, revealed strikingly consistent abnormalities, even in patients without obvious epilepsy. As early as 18 months of age, the background activity during wake periods slows and loses spatial differentiation. Between the ages of 3 and 8, the EEG voltage decreases and the records flatten. The waves become increasingly disorganized over time, eventually acquiring markedly abnormal activity during sleep. The contrast between the sleep and wake records is consistent enough to suggest the syndrome once children have reached that stage (reviewed in Hagberg, 1989Hagberg B.A. Pediatr. Neurol. 1989; 5: 75-83Abstract Full Text PDF PubMed Scopus (81) Google Scholar). It didn't give us an inroad to pathogenesis, however. The occurrence of Rett syndrome almost exclusively in females suggested a genetic basis for the disorder, despite the fact that 99% of cases are sporadic. Hagberg and colleagues proposed that an X-linked dominant mutation would likely produce affected females and non-viable hemizygous males (Hagberg et al., 1983Hagberg B. Aicardi J. Dias K. Ramos O. Ann. Neurol. 1983; 14: 471-479Crossref PubMed Scopus (1164) Google Scholar). More evidence in support of a genetic etiology came from very high concordance in monozygotic twins and discordance in dizygotic twins. But in the days before whole-exome and whole-genome sequencing, finding the culprit gene was a laborious process, even though we were pretty sure it had to be on the X chromosome. Since searching for genes in the early 1980s and 1990s was still daunting, we hoped to find patients with chromosomal abnormalities, which we could use to narrow down the candidate region on the X chromosome. Early studies from different labs focused on a possible fragile site in Xp22 in Rett, but that turned out to be common in the general population. That was the first red herring in Rett genetics. Two families each had one pair of affected half-sisters, with the mother the common parent, supporting the notion of X-linked dominant inheritance. We hypothesized that the mother carried a mutant allele but is protected because of favorable X chromosome inactivation, and this proved to be the case for one of the mothers. These two families helped us narrow the candidate region to the distal third of the X chromosome based on the pairs of half-sisters sharing this region of their maternal X chromosome (Ellison et al., 1992Ellison K.A. Fill C.P. Terwilliger J. DeGennaro L.J. Martin-Gallardo A. Anvret M. Percy A.K. Ott J. Zoghbi H. Am. J. Hum. Genet. 1992; 50: 278-287PubMed Google Scholar). We thought we had our lucky break when we identified a patient with Rett and an X:3 translocation, but this turned out to be the second red herring: cloning the break point revealed that no genes were disrupted by the translocation. Hope rose again when we identified a family with two second half-cousins with Rett syndrome who are related through maternal lines and share a common maternal great-grandmother. Exclusion mapping, however, showed that these two affected cousins shared no parts of the great-grandmother's X—a huge disappointment and one that almost killed the X-linked hypothesis. We identified a patient with an inversion on the X chromosome, but here again the inversion breakpoint in Xq26 did not yield the gene—the third red herring. Then, Schanen and Francke identified a family with an affected aunt and a niece, which helped narrow the region on Xq a bit more (Schanen et al., 1997Schanen N.C. Dahle E.J. Capozzoli F. Holm V.A. Zoghbi H.Y. Francke U. Am. J. Hum. Genet. 1997; 61: 634-641Abstract Full Text PDF PubMed Scopus (96) Google Scholar), and exclusion mapping data from a fourth family studied by Hoffman and colleagues further narrowed the region to Xq28 (Sirianni et al., 1998Sirianni N. Naidu S. Pereira J. Pillotto R.F. Hoffman E.P. Am. J. Hum. Genet. 1998; 63: 1552-1558Abstract Full Text Full Text PDF PubMed Scopus (145) Google Scholar). In 1999, after 16 years of mapping and systematic sequencing of candidate genes, we finally discovered the cause of Rett syndrome (Amir et al., 1999Amir R.E. Van den Veyver I.B. Wan M. Tran C.Q. Francke U. Zoghbi H.Y. Nat. Genet. 1999; 23: 185-188Crossref PubMed Scopus (3802) Google Scholar): loss-of-function mutations in the gene encoding Methyl-CpG binding protein 2 (MECP2). All the girls that helped us narrow the candidate region on the X chromosome had loss-of-function MECP2 mutations. The girl with the inversion had a MECP2 mutation on the non-inverted X chromosome (the inversion was coincidental). The two half-cousins that did not share any DNA did indeed have MECP2 mutations, but the disorder was sporadic in both girls—the odds of this happening twice in the same family are vanishingly small. But rare events happen, and these "red herrings" are a reminder that, even in this day of whole-genome interrogations, sequence abnormalities in patients must be carefully interpreted in the context of all possibilities and the rest of the genome. Fishing expeditions, so often maligned by grant reviewers, can be abundantly fruitful. After the Rett gene expedition, more fishing paid off. We began to appreciate that many syndromes in the broader classes of intellectual disabilities (IDs) and autism spectrum disorders (ASDs) can, like Rett, be both sporadic and genetically determined. ID and ASD are highly heterogeneous, however, whereas Rett syndrome is a single-gene disorder, which facilitates pathogenesis and mechanistic studies. Today, we know that ∼95% of individuals with classic Rett syndrome have mutations in MECP2 (the remaining 5% may have mutations in regulatory control regions, given that sequencing has thus far focused on the coding region). The discovery of the gene expanded studies beyond typical Rett syndrome to atypical versions: girls with either mild hypomorphic mutations, favorably skewed X inactivation patterns, or both can present with only a subset of features—e.g., mild ID and obsessive-compulsive behaviors without midline hand movements (Figure 1)—or even simply with slight cognitive deficits or autism. MECP2 mutations that would produce classic Rett in females cause males to suffer neonatal encephalopathy and die in infancy; milder mutations such as the A140V missense mutation allow boys to survive, albeit with intellectual impairment and sometimes motor or psychiatric disturbances as well (Figure 1). The more common MeCP2-related disorder to occur in males comes from an extra copy of MECP2 through duplications spanning the locus, which causes MECP2 duplication syndrome (MDS) (Van Esch, 2012Van Esch H. Mol. Syndromol. 2012; 2: 128-136PubMed Google Scholar). Although the smallest region of overlap spans MECP2 and IRAK1, the finding that doubling MeCP2 levels in mice reproduces all features of the human disorder pinpoints MECP2 as the culprit. On paper, as MDS unfolds, it looks a lot like Rett syndrome, with its absent speech, stereotyped behaviors, loss of purposeful hand use, delayed ambulation, and seizures. However, the problems start much earlier and include hypotonia from birth, which causes delayed sitting, standing and walking; there is also a tendency to contract severe respiratory infections. Females with the duplication can be asymptomatic carriers or have mild psychiatric symptoms, such as depression and obsessive compulsive disorder (OCD), or developmental delay with milder versions of the phenotype. As would be expected, triplications spanning the MECP2 locus in humans or tripling the levels of MeCP2 in mice cause a more severe phenotype. In 1992, in the laboratory of Adrian Bird, MeCP2 was originally discovered as a factor that binds methylated cytosines followed by guanine (CpG) (Lewis et al., 1992Lewis J.D. Meehan R.R. Henzel W.J. Maurer-Fogy I. Jeppesen P. Klein F. Bird A. Cell. 1992; 69: 905-914Abstract Full Text PDF PubMed Scopus (1077) Google Scholar) and later found to bind all methylated cytosines (reviewed in Lombardi et al., 2015Lombardi L.M. Baker S.A. Zoghbi H.Y. J. Clin. Invest. 2015; 125: 2914-2923Crossref PubMed Scopus (148) Google Scholar). It is highly abundant in the brain (Ross et al., 2016Ross P.D. Guy J. Selfridge J. Kamal B. Bahey N. Tanner K.E. Gillingwater T.H. Jones R.A. Loughrey C.M. McCarroll C.S. et al.Hum. Mol. Genet. 2016; (Published online August 9, 2016)https://doi.org/10.1093/hmg/ddw269Crossref Scopus (45) Google Scholar), where its function seems to be critical for most neurons and glial cells. While we still do not have a clear picture about the exact molecular cascade by which MeCP2 dysfunction leads to neuropsychiatric phenotypes, we do know that loss or gain of MeCP2 leads to thousands of modest gene expression changes (reviewed in Lombardi et al., 2015Lombardi L.M. Baker S.A. Zoghbi H.Y. J. Clin. Invest. 2015; 125: 2914-2923Crossref PubMed Scopus (148) Google Scholar). With such broad effects, it has been challenging to find a simple target or pathway that explains the disorder. The multiple mouse models, however, have provided insights into various aspects of the biology of the disease, and conditional deletions of murine Mecp2 from different groups of neurons link particular neuronal populations with particular symptoms (Lombardi et al., 2015Lombardi L.M. Baker S.A. Zoghbi H.Y. J. Clin. Invest. 2015; 125: 2914-2923Crossref PubMed Scopus (148) Google Scholar). Mecp2 null mice develop a severe version of the disease, but re-introducing Mecp2 in adulthood rescues the phenotype, suggesting that the brain is sufficiently intact as to be treatable if and when we find viable therapies (Guy et al., 2007Guy J. Gan J. Selfridge J. Cobb S. Bird A. Science. 2007; 315: 1143-1147Crossref PubMed Scopus (876) Google Scholar). Various models bearing human mutations have provided important insights into the functional domains of the protein (reviewed in Lombardi et al., 2015Lombardi L.M. Baker S.A. Zoghbi H.Y. J. Clin. Invest. 2015; 125: 2914-2923Crossref PubMed Scopus (148) Google Scholar). We have also learned that some mutations reduce MeCP2 levels, which underscores the sensitivity of the brain to levels of this surprisingly abundant protein. MeCP2 is much more abundant in the cortex than in the cerebellum (Ross et al., 2016Ross P.D. Guy J. Selfridge J. Kamal B. Bahey N. Tanner K.E. Gillingwater T.H. Jones R.A. Loughrey C.M. McCarroll C.S. et al.Hum. Mol. Genet. 2016; (Published online August 9, 2016)https://doi.org/10.1093/hmg/ddw269Crossref Scopus (45) Google Scholar), which dovetails with the earlier neuropathological study showing the cerebellum to be much less affected (Armstrong, 2005Armstrong D.D. J. Child Neurol. 2005; 20: 747-753Crossref PubMed Scopus (176) Google Scholar). This also fits with Andreas Rett's observation that the gait abnormality is more of an apraxia rather than an ataxia. One of the most perplexing features of Rett syndrome is the delayed onset. MeCP2 binds methylated cytosines, both CG (mCG) and non-CG (mCH); mCH marks increase postnatally, so it could be mCH binding that is important (see Lombardi et al., 2015Lombardi L.M. Baker S.A. Zoghbi H.Y. J. Clin. Invest. 2015; 125: 2914-2923Crossref PubMed Scopus (148) Google Scholar). There may be more to the delay than non-CG methylation, however. Animals with adult deletion of Mecp2 take some time to manifest the phenotype, but this period gets shorter if the gene is manipulated in older animals (Du et al., 2016Du F. Nguyen M.V. Karten A. Felice C.A. Mandel G. Ballas N. Hum. Mol. Genet. 2016; 25: 1690-1702Crossref PubMed Scopus (19) Google Scholar). Likewise, there is a similar delay for rescue: when MeCP2 levels are normalized in the duplication mice, there is a lag of about 4 weeks before gene expression and behavior normalize (Sztainberg et al., 2015Sztainberg Y. Chen H.-M. Swann J.W. Hao S. Tang B. Wu Z. Tang J. Wan Y.-W. Liu Z. Rigo F. Zoghbi H.Y. Nature. 2015; 528: 123-126Crossref PubMed Scopus (8) Google Scholar). Whether it takes time for chromatin architecture or network plasticity to deteriorate or recover or the cascade of gene expression changes to unfold is unknown. Another puzzling feature of Rett syndrome is the order in which symptoms appear. Regardless of the age at which normal development stops, there are 2 to 6 months of stagnation, then a regression phase during which the child withdraws socially and emotionally and develops apraxia and stereotyped hand motions. Whatever language skills were once gained vanish by the end of the second year, and then seizures, spasticity, and breathing dysrhythmias appear. Decades later, survivors develop parkinsonism. The early vulnerability of cognitive and social function became most apparent to me when I saw a family with the aforementioned A140V mutation. Both brothers presented with mild intellectual disability, attention deficit and hyperactivity, aversion to change, and aggression; one of them developed epilepsy. Their mother told me that her two brothers had a disorder similar to that of her sons but that they developed motor problems and spasticity in their 30s, became wheelchair bound in their 40s, and developed breathing problems and died in their 50s. While we don't have DNA to confirm that the uncles carried the A140V mutation, their clinical picture is strongly suggestive and consistent with some of the cognitive and motor features seen in other individuals with the same mutation (Lombardi et al., 2015Lombardi L.M. Baker S.A. Zoghbi H.Y. J. Clin. Invest. 2015; 125: 2914-2923Crossref PubMed Scopus (148) Google Scholar). It is fascinating that the panoply of symptoms occurs in the same order even when it is drawn out over decades; it's also striking that cognition, mood, and social interaction deficits are invariably the first signs. Recall that girls with mild mutations or favorably skewed X inactivation may present only with mild learning disability. It therefore seems that some neurons or circuits are more vulnerable to MeCP2 dysfunction than others. The vulnerability of neuronal circuits involved in cognition and social behavior might explain why mutations in so many different genes can lead to ID and ASD. It is conceivable that mutations causing classic nonsyndromic autism, bipolar disorder, or schizophrenia without ID result from hypomorphic alleles or even multiple common variants of genes that, with null alleles, cause syndromic ID, ASD, and schizophrenia. Loss of MeCP2 leads to expression changes in thousands of genes, compromises the majority of brain cells and circuits, and dysregulates all neurotransmitter systems. Developing a treatment will be no small feat. No study thus far has identified a protein that can substitute for MeCP2 function. Efforts to activate the silent wild-type allele or to deliver MeCP2 via gene therapy are ongoing but face significant challenges such as confining the activation to MECP2 on the silent X and ensuring broad nervous system delivery at the desired levels for gene therapy. For children with mutations that reduce MeCP2 levels, it might be possible to upregulate MeCP2 expression. In the case of MECP2 duplication disorder, lowering MeCP2 levels using strategies that target its regulators or anti-sense oligonucleotides should be feasible (Sztainberg et al., 2015Sztainberg Y. Chen H.-M. Swann J.W. Hao S. Tang B. Wu Z. Tang J. Wan Y.-W. Liu Z. Rigo F. Zoghbi H.Y. Nature. 2015; 528: 123-126Crossref PubMed Scopus (8) Google Scholar). Modulating particular circuits through the use of deep-brain stimulation, near-infrared light, or some yet-to-be-developed chemical or optogenetic approach is another possibility. Regardless of the strategy, there will be two major considerations for treatment trials. The first is timing. It is more likely for a trial to show a benefit within a reasonable time period if the phenotypes are targeted early in development. Once we demonstrate benefits before symptoms are set, we can always design longer-term trials for the older and chronically symptomatic individuals. The second is the need to combine a trial with appropriate behavioral, physical, and cognitive therapies. Given that individuals with Rett have not been through the typical language development and learning that healthy children experience early in life, a trial should not examine the effect of a drug or other therapeutic modality in isolation from interactive cognitive and physical training, lest potentially effective treatments could be dismissed prematurely. When we reflect on the 50-year history of Rett syndrome, the importance of scrupulous history-taking and careful clinical observation over a long period of time are abundantly clear. Not only were they crucial to Andreas Rett and Bengt Hagberg when defining the syndrome as a distinct clinical entity, but without such detailed description, we would not have recognized that a sporadic disorder could have a genetic basis. The history of Rett syndrome also teaches us that sometimes "fishing expeditions" are necessary—and that nature rarely provides an easy hook. The discovery of the Rett syndrome gene and the ongoing partnerships between the lab and clinic (and close observation of myriad mouse models) allowed us to uncover the extraordinary range of disorders involving MeCP2 and has led to exciting studies bridging epigenetics and neurobiology. Last but not least, the reversibility of symptoms in the mouse models gives us hope that, as we continue to elucidate the biochemical and neural circuit abnormalities of these disorders, we have all the tools needed to develop viable therapies. I want to thank Kathy Hunter (International Rett Syndrome Association, now known as International Rett Syndrome Foundation) for her early work raising awareness and supporting research and Monica Coenraads (Rett Syndrome Research Trust) for directly motivating so much Rett research. More recently, the parents of MECP2 duplication syndrome children have lent their support. The tireless efforts of thousands of families affected by Rett and MDS to raise funds for research while caring for their children is inspiring. I am also grateful to my colleagues Vincent Riccardi, Alan Percy, Dawna Armstrong, and Daniel Glaze for all their clinical mentorship and contributions to the Rett patients at the Blue Bird Circle Clinic at Baylor College of Medicine and Texas Children's Hospital and, last but not least, to all members of the Zoghbi lab for their research efforts on Rett syndrome and MECP2 disorders, especially Ruthie Amir for her extraordinary tenacity.