Skewed X-chromosome inactivation is not associated with premature ovarian failure in a large cohort of Italian patients^†

To the Editor:

Premature ovarian failure (POF) is a heterogeneous disorder defined as cessation of menses before the age of 40 years with increased levels of serum gonadotropins. POF is a common disorder, affecting approximately 1% of females in the Caucasian population and it can be caused by either genetic or environmental factors. Several familial cases have been described but the model of inheritance is not fully understood and only a small fraction of cases seems to be caused by mutations in the autosomal genes identified to date [Goswami and Conway, 2005]. The FMR1 premutation in Xq27 is the most common risk factor for POF, and it underscores a role for the X chromosome in the etiology of the disorder. A role for X-chromosome genes was suggested also by the relatively large number of structural abnormalities found in POF patients: primary amenorrhea and streak ovaries are associated with X-monosomy, while POF of variable severity is associated with partial X-chromosome monosomies and other rearrangements [Bione and Toniolo, 2000]. The role of the X-chromosome genes is, however, not well defined, as most of the genes involved have not been identified. The problems encountered in their identification suggest that they do not play a major role in the etiology of POF [Bione et al., 2004].

In this respect, the results of Sato et al. 2004, who reported a significant association of skewed X-chromosome inactivation (XCI) and POF in a group of patients from Japan were surprising as they suggested that POF may be frequently caused by genetic defects inducing skewed XCI, namely alterations of X-chromosome genes.

We evaluated the degree of X-inactivation in a large cohort of POF patients from Italy, presenting with secondary amenorrhea (age at menopause: range 12–39; mean ± SD: 32.5 ± 6.2). Since blood samples were collected within 6 months from the time of the diagnosis, age and age at menopause were substantially equal. In all cases, a normal 46;XX karyotype was demonstrated by standard banding techniques on metaphase chromosomes. A total of 151 POF samples and a control group of 164 age-matched normal women (age: range 23–42; mean ± SD: 32.9 ± 4.3) informative for the XCI test were studied. The degree of XCI skewing was determined by amplification of the AR locus before and after digestion with the methylation-sensitive HpaII restriction enzyme [Allen et al., 1992]. Complete digestion was monitored by amplification of the MIC2 locus [Anderson and Brown, 2002]. After electrophoresis on a MegaBACE 1000 DNA Analysis System (Amersham Biosciences, Munich, Germany), the peak area of the alleles was determined by the Genetic Profiler Analysis software (Amersham Biosciences). Skewing was calculated as described in Beever et al. 2003. Samples presenting >80% skewing were analyzed two to three times and the average of the different determinations was calculated.

A large number of women in both groups presented with skewing of XCI. When we classified them under four different thresholds (≥70, 80, 90, and 95%), comparable frequencies were found among POF patients and control women (Table I) in all groups. About 6% of the women had extreme skewing (>90%), a value similar to the 3–7% reported for healthy females in the same age range [Busque et al., 1996; Gale et al., 1997; Uehara et al., 2001; Beever et al., 2003; Hatakeyama et al., 2004]. In our sample we were able to distinguish a very extreme threshold presenting ≥95% skewing: in this group the number of POF patients was twice that of the controls. The difference was not statistically significant (X² test, 1 d.f., P = 0.185) and the small size of the sample may be responsible for the difference.

Among the POF patients analyzed, six were carriers of the FMR1 premutation: they did not show evidence of preferential skewing of XCI (respectively, 53, 61, 64, 66, 69, and 69%).

Our analysis showed that the distribution of skewed XCI in the Italian POF population was very similar to that determined in the control population and to previous data [Hatakeyama et al., 2004]. Our results are in contrast to what reported by Sato et al. 2004 who observed a significant association of skewed XCI to the POF condition. The etiology of POF in the Japanese population may be different, and the frequency of POF is lower than among Caucasians [Luborsky et al., 2003]. However, the discrepancy with our data could be more likely ascribed to the small size of the sample analyzed by Sato et al.: in the control group of 29 individuals they were unable to find women with extremely skewed XCI (>90%), which were previously indicated by the authors as the 5.6% of normal Japanese women [Uehara et al., 2001]. Lack of this group of females mainly contributes to the statistical significance of the report.

Moreover, the age range of the POF patients (range: 22–50) in the Sato group suggests that at least some of the patients were recruited long after the onset of premature menopause. If this is the case, increased levels of skewed XCI in this group could be due to menopause-related hormonal changes that have been proposed as one possible explanation of the well-documented drastic increase of skewed XCI in women older than 60 [Hatakeyama et al., 2004; Kristiansen et al., 2005].

In conclusion, skewed inactivation does not appear to be a common feature among POF patients and genetic defects able to induce preferential inactivation of the abnormal X chromosome (i.e., small deletions or mutations in X-linked genes) do not appear to be a common feature of POF patients. Thus, the X-linked genes that are likely to be involved in the disorder may be too few or not able to interfere with XCI in a population based assay, as it seems to be the case for the FMR1 premutation.