Response: To be skewed or not to be? This is no longer the question

In their letter to the editor, Dr Prchal's group raises several issues. We agree with some of these issues and respectfully disagree with others.

First, the T_m for the 3 IDS primers used by both our groups varies between 51.6°C and 52.9°C (not 60°C). Using annealing temperatures between 45°C and 55°C, we observed identical Ct values and excellent discrimination between IDS alleles, whereas an annealing temperature of 60°C resulted in less efficient amplification. Furthermore, we obtained identical results using the carboxyfluorescein-labeled IDS-specific probe (used by Swierczek et al¹ ) and SYBR green. Thus, as reported,²  the quantitative allele-specific polymerase chain reaction transcription assay (TCA) that we used was not identical, but very similar, to that used by Swierczek et al.¹ 

The incidence of skewing doubles from birth to age 30,^3,4  and reaches 40% in the blood cells of 60-year-old women.⁴  Our cohort was composed of relatively old subjects (mean age of subjects < 65 years was 59 years), hence our high incidence of skewing. Had we used the 80:20 criteria of skewing used by Swierczek et al,¹  skewing incidences would have been 29%, 31%, and 31% using HUMARA, quantitative TaqMan single nucleotide polymorphism TCA, and allele-specific polymerase chain reaction TCA, respectively. The skewing incidence observed by Swierczek et al¹  using HUMARA was similar, suggesting that groups in this age range are comparable.

The fundamental issue is whether the skewing observed in blood cells of elderly women is an artifact caused by acquired methylation modifications at the HUMARA locus¹  or a genuine biologic phenomenon. In our study, we used 2 independent transcription clonality assays (one based on allele-specific priming, the other based on quantitative TaqMan TCA) to analyze X-inactivation at the same locus (IDS), and found a high incidence and concordance of skewing. The highly consistent results obtained using 2 orthogonal transcription clonality assays argue strongly for the robustness of our findings and conclusions. The concordance of these results with those obtained by HUMARA further demonstrates that skewing is loci- and assay-independent.

Concordance between transcription-based assays and HUMARA does not rule out the possibility that age-dependent acquired epigenetic changes might affect assessment of X-chromosome skewing.⁵  Such alterations could affect both methylation- and transcription-based assays. However, it is unlikely that age-dependent methylation modifications are the primary basis for age-dependent skewing because they would have to: (1) be highly prevalent to explain the approximately 30% incidence of skewing; (2) preferentially occur in the maternal or paternal X chromosome in all cells of the woman; and (3) occur similarly at multiple, unlinked loci (HUMARA and PGK⁴ ; HUMARA and IDS² ).

We believe the issue is not which assay is best or whether the incidence of skewing increases with age in blood cells. Rather, the key remaining question relates to the etiology of nonrandom X inactivation in aging women. We suspect it is multifactorial, possibly influenced by epigenetic changes, and determined at least in part by hemizygous cell selection occurring over time.⁶  Although we⁴  have not implied that skewing is an a priori indicator of clonal hematopoiesis, this remains an etiologic possibility, at least in some persons. Genetic proof of clonal hematopoiesis will require the identification of acquired somatic mutations in patients with X-inactivation skewing, as has been demonstrated for myeloproliferative neoplasms.⁷  We look forward to subsequent studies into X-inactivation, as this remains an important area of investigation with many unanswered questions.