In heterozygous females, an unbalanced X chromosome inactivation pattern (skewed lyonization) may cause disease expression of X-linked disorders,1 for example, X-linked sideroblastic anemia (XLSA).2,3 X chromosome inactivation analyses such as the polymerase chain reaction (PCR)–based human androgen receptor assay (HUMARA) can reveal whether a female has a balanced or a skewed lyonization. Skewing itself can be constitutional or acquired for many different reasons.1 Moreover, in female carriers of X-linked disorders, skewed lyonization can be fortunate or unfortunate.1 The latter means a predominant inactivation of the X chromosome harboring the wild-type allele. But the distinction between skewed and balanced lyonization depends on various arbitrary definitions as well as certain technical variables.
Cazzola et al4 recently reported an Italian family with females heterozygous for an ALAS2 mutation that may cause XLSA. The degree of lyonization in these individuals was determined in leukocytes by the cleavage ratio between alleles from the HUMARA. For this assay, the methods for the detection and semiquantitative assessment of PCR products have crucial importance. Cazzola et al used silver-stained nondenaturing polyacrylamide gels and densitometric scanning for determination of cleavage ratios and did not provide methods for their calculation or correction. For semiquantitation, we recommend the use of an automated laser fluorescence sequencer or a similar device for enhanced resolution.5
Cazzola et al, like some other authors, attribute “excessive skewing”4(p4364) to allele ratios higher than 3.0 while allele ratios below 3 are defined as balanced. Elsewhere, a ratio between 1.85 and 4.0, as found in the 3 female carriers, has been termed “moderately skewed,”6(p32) and several other authors considered an “extreme lyonization” or “monoclonality” only when allele ratios were above 10.7(p62),8(p1581) The allele ratio can also be translated into the percentage of inactivated X chromosomes harboring the wild-type allele as follows: [ratio / (ratio + 1)] × 100.9 Thus, for case II-2 in Cazzola et al's report, the ratio of 3.2 would translate into 76% of cells with an inactive wild-type ALAS2 allele. But sequence analysis of cDNA derived from her reticulocyte RNA showed only expression of the wild-type allele. This finding is not discussed and is difficult to reconcile. Theoretically, a distinct erythroid lineage–specific X chromosome inactivation pattern (XCIP) due to a postinactivation selection may provide a possible explanation and could be resolved by X chromosome inactivation analysis of erythroid precursors.
Unfortunately, the only anemic person in the family reported by Cazzola et al (the proband; hemoglobin level 5.2 g/dL) was not informative, but an extremely skewed lyonization, for example, 99% of cells having an inactive wild-type ALAS2 allele, can be assumed. Using the above formula in cases II-3 and III-2 (with HUMARA cleavage ratios of 4.0), 80% of the cells should have an inactive wild-type ALAS2 allele. It is puzzling how 20% of cells with an active wild-type ALAS2 allele can account for the lack of anemia in these individuals.
Finally, the authors postulate familial skewing. But the moderately skewed XCIP in the 3 females' leukocytes could also be the result of an age-related stochastic event, as occurs in approximately 35% of normal females.7 A comparison of leukocyte XCIP with XCIP from other tissues is needed.
X chromosome inactivation ratios in female carriers of X-linked sideroblastic anemia
Aivado et al raise a number of questions concerning our recent paper on familial-skewed X chromosome inactivation as a predisposing factor for late-onset X-linked sideroblastic anemia (XLSA) in carrier females.1-1 We are pleased to provide them with technical details that could not be placed in a brief report and also to have the opportunity of discussing the pathophysiology of sideroblastic anemia.
Aivado et al correctly state that distinction between skewed and balanced lyonization depends on various arbitrary definitions as well as certain technical variables. They claim that we did not provide methods for calculation of cleavage ratios or their correction but do not consider that we had just one sentence available for describing clonal analysis of hematopoiesis. Therefore, we referred the reader to our previous methodological paper,1-2 which can provide technical and methodological details. It is a shame that our colleagues did not have the chance to read this article.
The decision to use a cleavage ratio equal to 3.0 as the cutoff between cases with balanced X chromosome inactivation and cases with excessive skewing was arbitrary by definition. More generally, any cutoff is arbitrarily established (eg, a hemoglobin level of 12 g/dL for distinguishing between healthy and anemic women): what counts is the rationale supporting the arbitrary decision. In the previously mentioned paper,1-2 we did perform a detailed analysis of the literature, which indicated that a value of 3.0 was the best cutoff. Our German and American colleagues recommend the use of an automated laser fluorescence sequencer or a similar device for enhanced resolution: we fully agree and have indeed started to use this technique in the last few months.
As regards case II-2 in our report, it is true that the ratio of 3.2 would translate into 76% of cells with an inactive wild-typeALAS2 allele [(3.2 × 100) / (1 + 3.2)]. Aivado et al find it difficult to explain the fact that sequence analysis of cDNA derived from her reticulocyte RNA showed only expression of the wild-type allele. They also argue that it is puzzling how 20% to 24% of cells with an active wild-type ALAS2 allele can account for the lack of anemia in women II-2, II-3, and III-2. What they do not account for is the pathophysiology of anemia in XLSA. We are glad to provide them with the interpretation of these findings that was included in the first version of our manuscript and eventually had to be omitted for reasons of space.
Despite the fact that our proband was not informative for clonal analysis of hematopoiesis, studies on the erythroid-specific 5-aminolevulinic acid synthase (ALAS2) structure and expression provided useful information. In fact, although she was heterozygous for the ALAS2 mutation, only the mutantALAS2 mRNA was expressed in her reticulocytes, as happened with her grandson, who is hemizygous and therefore carries only the mutant X chromosome. It should be noted that both the woman and her grandson were under pyridoxine treatment and no longer anemic at the time they were found to express the mutatedALAS2 allele. On the other hand, the remaining 3 heterozygous women in this family had normal hemoglobin levels and, despite unbalanced X chromosome inactivation, expressed the normalALAS2 in their reticulocytes. In the proband's daughters, red cell production is essentially sustained by erythroid cells carrying the nonmutant X chromosome as the active one. Even if such erythroid cells represent only about 20% to 24% of total immature red cells, they can clearly sustain a normal red cell production.1-3 Most erythroid precursors expressing the mutant ALAS2 are ring sideroblasts that die prematurely in the bone marrow, a mechanism responsible for anemia in hemizygous males and known as ineffective erythropoiesis. The few mature red cells deriving from erythroid precursors expressing the mutant gene account for the slightly increased red blood cell distribution width (RDW) values that are typically observed in heterozygous females. But the RNA content of reticulocytes expressing the mutant gene is only a small fraction of total reticulocyte RNA and may or may not be detected using the cDNA assay employed by us (which is semiquantitative). One experiment we did not carry out involved administration of pyridoxine to the nonanemic heterozygous women in order to see its effect on mutant ALAS2 expression in their reticulocytes. It is possible that under pyridoxine women II-2 and II-3 would also have expressed the mutant allele.
Finally, Aivado et al raise doubts about our conclusion that skewing was familial. They suggest that the moderately skewed X chromosome inactivation patterns (XCIPs) in the 3 females' leukocytes could also be the result of an age-related stochastic event. In our previous paper,1-2 we studied XCIPs in blood cells from healthy women belonging to 3 age groups: neonates (umbilical cord blood), women 25 to 32 years old (young women group), and women more than 75 years old (elderly women). The frequency of skewed X inactivation in polymorphonuclear cells (PMNs) increased with age: in fact, a cleavage ratio of at least 3.0 was found in 3 of 36 cord blood samples, 5 of 30 young women, and 14 of 31 elderly women. The inactivation patterns found in T lymphocytes were significantly related to those observed in PMNs in both young (P < .001) and elderly women (P < .01). Based on the above estimates, the probability that the 4 women in our family simply had age-related skewing would be 8 divided by 10 000 [(5 of 30)4], while the probability that skewing was familial is 9992 divided by 10 000. Consequently our conclusion had a strong scientific basis.
Aivado et al suggest that a comparison of leukocyte XCIP with XCIP from other tissues is needed. To define the best control tissue for the interpretation of X chromosome inactivation patterns in hematopoietic cells, we previously analyzed X chromosome inactivation patterns in different peripheral blood cell populations and in hair bulbs from healthy women belonging to different age groups.1-2 When PMNs were compared with hair bulbs,2(Fig3) no relationship was found with respect to the inactivation ratio (r = .31,P > .05). There was no difference between young and elderly women in this respect, a cleavage ratio of at least 3.0 in PMNs being associated with a similar value only in about 50% of hair bulb DNA from either young or elderly women.
In summary, findings of our study clearly indicate that the most likely explanation of the above findings is that the proband, despite a markedly congenitally unbalanced X chromosome inactivation in her hematopoietic cells, was able to produce normal amounts of red cells for the first 6 decades of her life, as her daughters and granddaughter do. In the seventh decade she developed acquired skewing, as do about one third of elderly women. She unfortunately further inactivated the parental X chromosome carrying the normal ALAS2 gene, and when nearly all red cell precursors expressed the mutant gene, she became severely anemic.
