RTT is a common monogenic cause of neurodevelopmental abnormalities in females [1–6]. Although it has been repeatedly noted that the phenotype of affected girls depends on the presence or absence of MECP2 mutation, the latter have not been ever considered as an exclusive criterion for RTT [11, 16, 17]. Apparently, non-striking phenotypic differences between a significant proportion of mutation-positive and mutation-negative cases  indicate that the same genetic defect causes the disease in mutation-negative cases. Currently, there have been reported several genomic abnormalities (i.e. 14q12 microdeletions) associated with RTT [5–7, 18, 19]. However, these genomic rearrangements are unlikely to cover all the mutation-negative RTT cases. Here, we report on the commonest cause of RTT in cases without detectable MECP2 mutations. To our knowledge, this is the first systematic report describing Xq28 genomic abnormities (Xq28 deletions affecting MECP2) in RTT.
Large intragenic MECP2 deletions have been consistently reported in the available literature [20–24]. Moreover, a RTT case was associated with a deletion detected by FISH . Nevertheless, the existence of Xq28 deletions causing RTT has long remained a matter of conjecture. It seems that the high mutation detection rate and clinical heterogeneity in mutation-negative cases has resulted in the lack of studies dedicated to whole genome analysis among RTT females without detectable MECP2 mutations. On the other hand, MECP2 loss modulates RTT phenotypes in mice [25, 26] suggesting that similar genomic abnormality might cause RTT in humans. Xq28 (MECP2) deletions found in RTT girls has confirmed this expectation. Furthermore, studying functional consequences of MECP2 mutations [27–29] evidences that MECP2 loss has functional implications in females.
As detected by array CGH and FISH, one deletion causing classic RTT was mosaic. Somatic mosaicism for a structural chromosome abnormality or CNVs is common in genomic disorders or single-gene disease [30, 31]. It is also detected in cohorts of individuals with autistic spectrum disorders (in its widest sense) including girls suffering from RTT [12, 18, 32–35]. This makes it attractive to analyze molecular and clinical aspects of Xq28 (MECP2) deletions in the light of increasing interest in biomedical studies of autism, especially considering the positive experience in modelling neurodevelopmental abnormalities according to data on RTT pathogenesis [36, 37]. To explain differences between cell proportions uncovered by array CGH and FISH (Figure 3), one can compare molecular cytogenetic techniques in context of detecting somatic mosaicism [38–40]. In this instance, we have concluded that FISH results are more accurate. Similarly, FISH questioned in some detail the size of the recurrent deletion causing classic RTT. Since oligonucleotide probes cover a part of MECP2 sequence whereas the deletion was detectable by FISH with a probe for MECP2, we have speculated that genomic loss within Xq28 is a bit larger than shown by the array CGH. Likewise, sequence variations specifically generating Xq28 subchromosomal rearrangements are co-localized with the breakpoints outside of MECP2 loci [41, 42]. So far, it appears to be also valid for reported deletions. To determine the intrinsic nature and causes of Xq28 (MECP2) deletions leading to classic RTT, further studies are indisputably required.
The specific replication patterns in RTT or type C (observed in about 90% of affected children in contrast to unaffected females [12, 43]) have been detected in females with Xq28 microdeletions. The type C replication pattern represents a disturbance in the sequence of replication in an inactive chromosome X apparently caused by MECP2 mutations [12, 15]. These data allowed speculations that RTT in mutation-negative females is likely to be associated with genetic defects affecting the MECP2 gene . Array CGH analysis of RTT girls, highlighting Xq28 (MECP2) deletion as a new cause of the disease, confirms this assumption.
Although RTT phenotype is characterized by recognizable patterns of malformation and distinct neurodevelopmental abnormalities, there does exist a clinical variability among females suffering from this severe disorder [3–6, 11, 14, 16, 17]. Xq28 deletions causing atypical RTT have shown to exhibit additional phenotypic features (Table 1). This can be easily explained, because all deletions have spanned significantly larger regions than the MECP2 locus, involving other Xq28 genes, as well (Figure 2). Conversely, Xq28 losses (MECP2 plus some additional genes) should naturally lead to the presence of phenotypic manifestations usually unseen in RTT. Interestingly, RTT females with large Xq28 deletions have demonstrated less severe disease manifestations as compared to their counterparts with intragenic MECP2 mutations of known functional consequences. This is likely to result from X chromosome inactivation skewing probably arisen from selective disadvantages of cells with an active deleted chromosome X. In the same way, MECP2 deletions causing classic RTT are likely to lead to less severe RTT manifestations through the skewed X chromosome inactivation patterns. Thus, epigenetic phenotype modulators determine the outcome of subchromosomal deletions involving MECP2. This has led us to the conclusion that, regardless of specific phenotypic appearance, the Xq28 deletion phenotype is not different enough from RTT due to intragenic MECP2 mutations to define it as an independent clinical entity or a microdeletion syndrome. Summarizing the clinical data on girls found to demonstrate Xq28 (MECP2) microdeletions, we have concluded that these genomic rearrangements cause at least two distinct RTT subtypes. The first subtype is caused by deletions spanning from 0.5 to 1 Mb and is characterized by less severe RTT manifestations as well as additional clinical signs. The second subtype is caused by deletions spanning about 100 kb leading to a loss of MECP2 per se and is simply a less severe classic RTT. Finally, both types can be arbitrarily designated as microdeletion RTT subtype.
To this end, it is to mention that submicroscopic genomic variations and CNVs are likely to be among the commonest causes of congenital malformations, idiopathic intellectual disability, autism, epilepsy, neuropsychiatric disorders [18, 36, 44]. Seemingly, these genome variations are likely to be important elements of pathogenetic cascades in complex disease mediating genetic-environmental interactions . The present study evidences that submicroscopic deletions or CNVs cause single-gene disorders in an appreciable proportion of cases.