We report the largest series of DMD symptomatic carriers, consisting of 24 cases covering a wide spectrum of clinical phenotypes. The main mechanism found to provoke symptoms in these patients was skewed XCI favouring expression of the DMD-mutated allele. We provide data consistent with the hypothesis that the extent of XCI skewing is related to phenotype severity and that skewed XCI patterns are in most cases familial inherited. We also describe what is, to our knowledge, the first report of DMD female carriers presenting cognitive impairment but no muscle weakness, broadening the clinical phenotypes associated with dystrophinopathy.
In line with previous reports
, isolated symptomatic carriers were relatively frequent, suggesting that dystrophinopathy should always be considered in females suffering from a myopathy of unknown cause, even in the absence of an X-linked family history. In these cases dystrophin immunostaining on muscle sections remains the best diagnostic tool even though an exon copy number screening technique (such as MLPA) can be considered prior to muscle biopsy. Muscle biopsies showed variable pathological features in all studied cases. Furthermore, dystrophin abnormalities were found in all but one case. In accordance with previous publications
[6, 20, 34], we did not find a correlation between dystrophin expression and clinical phenotype, although the most affected subjects showed generalized dystrophin absence (Figure
Our findings confirm that symptomatic carriers present a wide clinical variability ranging from nearly asymptomatic forms to disabling DMD-like phenotypes. The progression of muscle weakness among symptomatic carriers correlates with onset of symptoms. Onset during the first decade was associated in all cases with rapid progression leading to severe phenotypes (DMD-like or severe BMD-like), while later onset, from the third decade onwards, was associated with mild progression. However, patients presenting onset of symptoms during the second decade exhibited variable severity. Subject #7 presented onset of weakness at 20 years with rapid progression leading to wheelchair support five years later, whereas subject #11 who manifested the first symptoms at 18 years is still able to walk at the age of 49 years. In accordance with previous observations
[10, 12, 33], muscle weakness was asymmetric in 40% of our patients. Asymmetry may be caused by somatic mosaicism
 or reflect different XCI patterns between different muscle groups and tissues. In keeping with the later hypothesis, most of our cases presenting muscle weakness and random XCI in blood showed asymmetric distribution of weakness (4/5). Five patients presented symptoms without skeletal muscle involvement: two subjects presented isolated dilated cardiomyopathy while three showed cognitive abnormalities.
Cognitive abnormalities, observed in five patients in isolation or together with muscle weakness, ranged from behavioural issues to severe global developmental delay. Risk of cognitive deficits among DMD patients is thought to be the result of the cumulative loss of dystrophin isoforms in the central nervous system during development
. According to this hypothesis, mutations in three subjects (#5, #19 and #20) were predicted to destroy most dystrophin isoforms. Nevertheless, in the other two subjects (subjects #1 and #18) only long isoforms were predicted to be affected. It is of particular note that four of the five cases presenting cognitive deficits showed highly skewed XCI patterns in blood independently of muscle symptoms. This could indicate that the central nervous system is more affected than muscle by reduced dystrophin expression due to biased XCI.
Skewed XCI favouring expression of the DMD-mutated X chromosome has been proposed as the main mechanism accounting for symptoms in symptomatic carriers without chromosomal rearrangements. However, results concerning the use of XCI as a prognostic marker are controversial. While some studies found that the great majority of symptomatic cases exhibited skewed profiles
[20–22] others found that XCI is not a reliable measure to predict whether carrier female will develop symptoms
[6, 10, 24]. Most women in the normal population present a random XCI pattern in peripheral blood. However, a skewed pattern is present in 8.8% of females, increasing up to 14.2% in adult women
. We found a distribution of XCI ratios in blood among control groups (asymptomatic carriers and non-carrier females) similar to those described in the normal adult population
. Regarding our group of symptomatic carriers, excluding those manifesting only myalgia, the number of skewed cases was 49.2% higher (mean XCI ratio 18% higher) than in asymptomatic carriers. In cases presenting skeletal muscle involvement, the extent of XCI skewing in blood tended to increase in line with the severity of symptoms (Figure
3B). However, assessing XCI through the AR methylation assay seems to have a poor prognostic value, since some subjects showing similar XCI patterns exhibited different phenotype severity. Furthermore, four subjects (#16, #18, #19 and #20) with highly skewed patterns presented no skeletal muscle involvement. When XCI analysis was performed on muscle DNA we observed no correlation between the extent of XCI skewing and clinical severity or proportion of dystrophin-negative fibres. As previously described
[6, 23], in most cases XCI profile in muscle differed significantly from that found in blood. This could reflect a biochemical and genetic normalization process in skeletal muscle
[25, 37, 38], or tissue-specific differences in XCI or in the methylation status of the AR gene. In keeping with the latter hypothesis, in subject #3 the preferential expression of the DMD mutated allele correlated with the XCI ratio in blood but not with that in muscle (Figure
2). This case suggests that the methylation status of the AR gene in muscle may not reflect in all cases that of the DMD gene.
Soltanzadeh et al.
 found that DMD deletions or duplications were significantly more associated with skewed XCI compared with point mutations suggesting a correlation between XCI and mutation class. Analysing our data, we found that skewed XCI was more frequent in patients with point mutations (5/7, 71.4%) compared with those carrying deletions or duplications (7/15, 46.7%). However, these differences do not reach statistical significance in a Fisher’s Exact Test suggesting that the DMD mutation class is not involved in the development of X inactivation skewness.
We found evidence of familial XCI skewing in 78.6% of analyzed families suggesting that in many cases this may be under genetic control. X inactivation is a complex process restricted to early embryogenesis in which an X chromosome is silenced at random depending on the stable expression of the cis-acting XIST gene
. Skewed X inactivation may be caused primarily by preferential silencing of one specific X chromosome due to genetic factors, or by chance due to the very limited number of precursor cells present at the moment of inactivation
. In X chromosome defects, skewing can also be secondary during development due to post-inactivation selection. Selection tends to preserve gene dosage: in balanced X:autosome translocations the normal X is generally inactivated while in X deletions the inactivated chromosome is the deleted X
. Mutations in the XIST gene promoter have been reported to cause primary non-random X-chromosome inactivation
[31, 32]. We analyzed the XIST promoter in all patients but no changes were detected. Other loci on the X chromosome have been associated with familial skewing of X inactivation, suggesting that factors other than XIST may control the primary choice of X chromosome or be secondarily involved in cell survival
We consider that it could be clinically important to identify the genetic factors that alter random X inactivation in heterozygous females expressing X-linked recessive traits such as dystrophinopathy.
The reduced sample size and lack of systematic case collection limit the generalizability of the results.