Rett syndrome (RS; MIM 312750) is a severe neurological disorder affecting exclusively females.¹ Its prevalence is about 1 in 10 000 female births.² It is a prominent cause of profound mental handicap in women.³ The clinical course of the disease is typical, and consists of a normal neonatal period, followed by an arrest of development between 6 and 18 months of age. The patients show a number of clinical signs indicationg a neurodevelopmental defect: arrest of brain development, regression of acquired skills and behavioural problems (stereotypic hand movement, autism).⁴ The vast majority of cases are sporadic, although a few families have been reported.⁵ Mutations in the methyl CpG-binding protein 2 (MECP2) gene were identified in 1999 in females affected by RS,⁶ ending a relentless hunt that lasted for >15 years since the condition was brought to the attention of the medical community by Hagberg et al.¹ Soon after this major breakthrough, the first mutation in the MECP2 gene in a male patient was described.⁷ This mutation was identified in several individuals from a familial case of RS that had contributed to map the disease in Xq28.^5,8 In this unique family, two sisters were carriers of the G269fs mutation in MECP2. One of these women had two affected children: a girl affected by the classical form of RS, and a boy who died in early infancy from a severe neonatal encephalopathy of unknown origin. They were both carriers of the G269fs mutation. Until that date, due to the exclusive female occurrence of RS, it was believed that any mutation causing the disease led to early termination of putative male pregnancies. This is a classical situation for several X-linked dominant disorders (see Franco, Ballabio⁹ for a recent review). The fact that an excess of male miscarriages was not observed in the known familial cases of RS could be attributed to their exceptionally small number, and to the small number of children in these rare families. The discovery by Wan et al⁷ instantly revealed that the presence of a mutation in the MECP2 gene on the single X chromosome of a male embryo was compatible with development and life. This important and unexpected discovery prompted most laboratories involved in mutation screening for RS to include male cases in their screens. Today, it is clear that MECP2 mutations in male patients are not rare. Surprisingly, they are responsible for a wide spectrum of neurological disorders, ranging from mild mental retardation (MR) to severe neonatal encephalopathy. The aim of this review is to describe the nature of the MECP2 mutations identified in male patients (table 1) and their associated phenotypes (table 2). Cases reported between December 1999 and December 2006 are included. Non-pathogenic variants are not considered.

FREQUENCY OF MECP2 MUTATIONS IN MALES

MECP2 mutation screening was reported for 2697 male individuals affected by a neurological disorder. In this heterogeneous population of patients, 46 potentially disease-causing mutations were identified in different families and sporadic cases. They can be divided into two groups: 34 mutations that are certainly pathogenic (nonsense, frameshift and other mutations found in several girls with RS) and 12 unclassified variants consisting mainly of missense mutations found in a single family. The frequency of potentially disease-causing MECP2 mutations in the population of mentally retarded male patients is thus between 1.3% and 1.7%. This is an important figure, considering that the incidence of fragile X syndrome, the most frequent familial cause of MR in males, is 2.8% in the same population.⁴⁴

Five out of seven mutations identified by Del Gaudio et al^29,33are not taken into account in the numbers given above, as they were identified in a population of 1380 individuals whose phenotype and gender are not specified (these mutations are nonetheless listed in tables 1 and 2).

Many non-pathogenic nucleotide changes were also identified in the MECP2 gene. This gene has a very high rate of de novo mutations, and several polymorphisms were found in affected and healthy individuals in the same family.⁴⁵ These polymorphisms are not taken into account in the figures given here.

Frequencies can be calculated only if the number of screened patients is large enough. The first report of a large male population screened for mutations in MECP2 involved 185 mentally retarded patients negative for fragile X syndrome testing.¹² In total, four (2.1%) mutations were reported as disease causing in this cohort. This surprisingly high figure prompted other laboratories to screen more similarly selected patients. However, the results showed a much lower incidence of mutations. In the subsequent reports involving a total of 829 patients,^46,47 a single disease-causing mutation was identified (0.1%). Hence, testing negative for the expansion of the FMR1 CGG repeat does not seem to be a very useful criterion to select a population for MECP2 mutation screening.

After these misleading initial findings, subsequent screens were extended to include males affected by non-specific MR. In the 658 patients reported to date,^{21,22,48–52} 2 (0.3%) disease-causing mutations were identified. The first mutation was present in a two-generation family with three affected males,²² and the second in a male patient with unexplained MR.²¹

Targeted screens were also performed following the description of four patients with non-specific MR and the A140V mutation,^12,13 and five affected males in a single family with the psychosis, pyramidal signs and macro-orchidism (PPM-X) syndrome and the same A140V mutation.¹⁴ In total, 433 males with various forms of MR were screened for the presence of the A140V mutation,^52,53 but no mutation was found, questioning the real frequency of this amino acid change in the mentally retarded male population.

Because of the partial phenotypic overlap between RS and patients having a defect of the 15q11q13 region causing Angelman or Prader–Willi syndrome (PWS),^54,55 several studies tried to determine whether mutations could be found in MECP2 in patients negative for defects in this imprinted region. In a screen of 92 male patients negative for methylation defects at the UBE3A locus,^39,52,56,57 mosaic mutation was found in just 1 (1.2%) patient.³⁹ The patient in the report by Hitchins et al⁵⁷ with the P56fs mutation had already been described twice.^37,54 A G428S mutation described as pathogenic in such a patient⁵⁶ was shown to be a rare non-pathogenic variant.²⁰ Because the patient reported by Kleefstra et al²⁹ had a phenotype evocative of PWS, 71 male patients negative for PWS were also screened for MECP2 mutation but no mutation was identified.^39,52 A cohort of 154 male autistic patients were screened, but no disease-causing mutation, was found.⁵⁸ More recently, large duplications involving the MECP2 locus were identified in 18 male patients with a severe neurological phenotype.^30–33 These duplications represent 18 out of 34 (53%) of the currently known pathogenic mutations in this gene (18/34).

SCREENING METHODS

Different screening methods were used to search for mutations in the MECP2 gene. Direct sequencing was used most often because MECP2 is a small gene (1.5 kb of coding sequence). Denaturing gradient gel electrophoresis,¹² denaturing high-performance liquid chromatography⁴⁷ or single-strand conformation polymorphism screening⁵² were also used to screen large series of patients. These different techniques yield different mutation detection rates (between 0 and 3%). However, it is not possible to compare the yield because both the technique and the population that was screened are different (see above).

Since 2004, a new MECP2 exon has been known.⁵⁹ It is a small 5′ exon coding for 21 amino acids. Studies performed before this discovery did not include this small exon 1 in their screens. However, a large screen of 410 mentally retarded males originating from familial cases (with at least two affected individuals) for mutations in this exon of MECP2 found no mutation.⁶⁰ This indicates that mutations in exon 1 will not be a frequent cause of neurological disorders in male patients. It is also the case for RS, in which mutations in exon 1 account for only 1–3% of the known mutations.^61,62

Multiplex ligation-dependent probe amplification,^32,63,64 microarray-based comparative genomic hybridisation,^31,33 quantitative PCR³⁰ and dosage-sensitive Southern blots³³ were used to detect large rearrangements involving the MECP2 locus in mentally retarded males. These techniques allowed the detection of a large number of new mutations (17 duplications and one triplication).

POINT MUTATIONS FOUND IN MALES AND IN FEMALES WITH RS

The first group of mutations was identified in male patients, because they had a sister with RS. The first mutation in MECP2 was identified in a boy who died before the age of 1 year. He was a carrier of the G269fs mutation.⁷ He had a sister with the same mutation and a classical RS phenotype. Their mother was also a carrier of the same mutation, but was protected by a favourably skewed X-chromosome inactivation (XCI) pattern. The next cases were two brothers having severe neonatal encephalopathy who died before the age of 12 months and were carriers of the T158M mutation.¹⁶ Their mother was asymptomatic, although she was a carrier of the same mutation. She also had a totally skewed XCI. Since then, these two mutations (T158M and G269fs) have been found in boys with severe neonatal encephalopathy who did not have a sister affected by RS.^17,25 In these latter cases, the mothers were not carriers of the mutation.

Other mutations reported in boys were G252fs,²⁴ G163fs,²³ R133C¹⁰ and S134C.¹¹ All these boys had a sister having the classic form of RS, and they were all affected by a very severe neurological phenotype since birth. Two additional mutations were described in males affected by a severe phenotype: R270fs and G269fs.¹⁵ These two patients died before the age of 3 years. They all had respiratory insufficiency, microcephaly, limb rigidity and movement disorder. More recently, the L386fs mutation was identified in a boy and his unaffected mother.²⁷ The mother had a completely skewed XCI.

POINT MUTATIONS FOUND IN MALES BUT NOT IN FEMALES WITH RS

The A140V mutation has never been described in a girl with a classical RS phenotype. However, it was described several times in mentally retarded males. It was found in four severely retarded males from the same family.¹³ A female individual in this family is also a carrier of the mutation and has mild MR. The four affected males, aged 27–40 years, have a normal head size but severe MR. They have no history of regression after an initial normal development. The only constant features in affected individuals were spastic paraparesis and distal atrophy of the legs.⁶⁵ This A140V mutation was subsequently found in two males with non-specific MR¹² and in affected males from the MRX79 family mapped to Xq28.⁶⁶ In the MRX79 family, there is a large phenotypic heterogeneity, even among affected individuals. All carrier females have random XCI, which has prompted questions about the pathogenicity of this A140V mutation.²⁰ The A140V mutation was also described in a 12-year-old boy with developmental receptive language disorder and childhood-onset schizophrenia.⁶⁷ The mutation was present in the patient’s unaffected mother, whose XCI pattern was not reported.

The study by Couvert et al¹² also described mutations in MECP2 in three families with non-specific MR mapping to Xq28. The first family had an in-frame deletion of 80 amino acids (P387del80). Affected patients have severe to mild non-progressive MR, with better motor skills than verbal ability.⁴³ The second family carries the R167W mutation. The patients have essential tremor with mild and non-progressive MR, poor motor coordination and difficulties with written language. The third family carries the E137G mutation. The patients have mild to moderate MR, poor speech articulation and, in some patients, verbal stereotypies with suspicion of regression of language skills in three. The R167W and E137G mutations are currently listed as unclassified variants, since there is no strong argument to favour pathogenicity.

Sporadic cases with mutations in MECP2 were also reported by Couvert et al.¹² The P399L and R453Q mutations were found in two patients with moderate to severe MR. Unfortunately, parental DNA was not available for these two cases. Since its description, the P399L mutation has been shown to be a polymorphism.²⁰ This same paper demonstrates that the G428S mutation reported in a patient with moderate MR⁵⁶ is also a non-pathogenic variant. These two examples called for caution with the interpretation of de novo missense mutations in MECP2. For this reason, most missense mutations in table 1 are listed as unclassified variants, unless they were found in several unrelated cases of RS in female patients.

The R344W mutation was reported in a male patient with an RS-like phenotype with no clinical details.²⁰ It is maternally inherited, but the XCI pattern of the mother was not specified. It is thus another unclassified variant. Ventura et al¹⁹ described a 6-year-old boy with the P322S mutation, which was also present in his mother. This patient has moderate MR with autistic features and epilepsy. A de novo P225L mutation²¹ was found in a patient with severe MR, spastic tetraplegia, dystonia, complete apraxia, neurogenic scoliosis and breathing irregularities.¹⁸ The same study reported the P405L mutation in a male individual with MR, epilepsy and autism. This mutation was also found in the mother and the sister of this patient who had borderline IQ. Unfortunately, XCI ratios were not described for these women. The F157I and K417M mutations were described in two patients with a severe neurological phenotype.¹⁵ The mother of the patient with the F157I mutation had several spontaneous abortions before having this affected child, raising the suspicion of a potential germinal mosaicism. The K417M mutation is inherited from a clinically normal mother who has a random XCI pattern in her blood cells. Its pathogenicity is thus questionable.

Pathogenicity is more obvious for frameshift or nonsense mutations. A Q406X mutation was identified in an Xq28-linked MR family.²⁸ Two severely retarded males and two mildly affected females are carriers. The females have a random XCI profile. A 1415del2 de novo mutation was described in a 10-year-old boy with moderate MR, obesity and gynaecomastia.²⁹ An 816dup7 mutation was found in a male patient said to have a classical RS phenotype.²⁶ However, signs of encephalopathy were visible from birth. He never crawled, never spoke and shows no stereotypic hand movements similar to the ones seen in female patients.

LARGE DUPLICATIONS

An unexpected number of Xq28 duplications involving the MECP2 gene have been described in male patients with severe MR. The first study reporting such a case described a 430 kb duplication in a boy with severe MR and features of RS.³⁰ Soon after, four duplications with sizes ranging from 400 to 800 kb were characterised in patients with a similar phenotype consisting of infantile hypotonia, recurrent respiratory infections and severe MR.³¹ Friez et al³² used MLPA amplification of the MECP2 gene to identify six additional families in which a large duplication of the MECP2 gene was present. Two MECP2 duplications were described recently,⁶⁸ but they have already been reported in a previous paper.³¹ Del Gaudio et al³³ described six patients with a duplication of MECP2 and one patient with a triplication of this locus. The triplicated patient has the most severe phenotype. However, in the report by Del Gaudio et al,³³ the criteria to select patients for screening are not described. The duplications in Xq28 usually involve several genes, and very often involve L1CAM. When they are transmitted, carrier mothers have a totally skewed XCI pattern favouring the expression of the normal X chromosome.

MALES WITH ABNORMAL KARYOTYPES

In a very small number of cases, a classical RS phenotype was described in male patients. The first two cases were two boys with Klinefelter syndrome (47,XXY) and the Y141X³⁴ or T158M mutation.³⁵ In this latter case, the patient had somatic mosaicism, since only 76% of his cells were XXY (see below). This case should remain exceptional, since the risk of being affected by Klinefelter syndrome (1/1000 male births) and to carry a disease-causing mutation in MECP2 (1/10 000) is 1/10 000 000.

Another male patient had a female 46,XX karyotype, but he was found to carry a copy of the SRY sex-determining gene attached to one of his X chromosomes. In addition to this rare event, he was also a carrier of the E455X mutation in MECP2.³⁶ His motor development was delayed, he could not speak at 24 months, and had truncal muscular hypotonia, microcephaly and spasticity. He lost purposeful hand skills at 6 months of age, and a deceleration of head growth at about 7 months of age.

SOMATIC MOSAICISM

The first case of somatic mosaicism was reported in a patient with a P56fs mutation.³⁷ He was presenting with a classical RS phenotype. Soon after, another boy with a 47,XXY population of cells in his muscle lineage was shown to carry the R270X mutation.⁶⁹ His blood and skin cells had a normal karyotype. He was also affected by a classical form of RS. The same R270X mosaic mutation was reported in an unrelated male patient with typical RS.⁴⁰ Two other male patients were shown to carry the mosaic R133H and T158M mutations.^38,39 Both of them have the classical RS phenotype.

CONCLUSION

The phenotype of girls with RS has been perfectly described, and consensus inclusion and exclusion criteria were adopted and revised on several occasions.^70–73 They include the description of variant cases. The situation is not so clear concerning males with mutations in the MECP2 gene. However, 7 years after the description of the first mutation in a boy with a severe phenotype, it is now possible to make a number of genotype/phenotype correlations and to distinguish the three groups of patients. The first group is composed of patients who have a mutation which is also found in typical cases of RS. These boys have severe neonatal encephalopathy, and they usually die in their first year of life. These mutations can lead to a milder phenotype (and to clinical RS) when they are diluted among normally expressing cells. This is the case of XXY boys, or when somatic mosaicism is present. The second group of patients has mutations that are not found in females with RS. These mutations are usually compatible with life into adulthood. The neurological presentation ranges from severe to mild non-specific MR. The third group is composed of males having a duplication of the whole MECP2 gene (and sometimes genes in its vicinity). The primary clinical features associated with this microduplication are non-specific, but they comprise a severe phenotype. Affected individuals experience infantile hypotonia, recurrent respiratory infection, severe MR, absence of speech development, seizures and spasticity. The recurrence of respiratory infections may be a criterion to distinguish these cases from other syndromes, as they occur in a context of normal growth.

In conclusion, MR caused by mutations in MECP2 is not rare in male patients. With an estimated frequency of 1.3–1.7% in males with moderate or severe MR, it must be considered when neurological, metabolic, genetic, biochemical, electrophysiological and imaging investigations are not informative.