The scale of the phenomenon is striking. In a landmark cross-biobank epidemiological study encompassing 883,574 female participants from eight biobanks across Europe and East Asia, 105,286 women—11.9 percent—carried detectable LOX in their peripheral blood DNA, and 5,475 women carried expanded LOX, defined as the mutation present in more than 5 percent of circulating leukocytes. Prevalence rises continuously with age: in the Mass General Brigham Biobank, detectable LOX increases linearly from 3.5 percent in women under 40 to 33.5 percent in women aged 80 to 85, before dropping to 26.3 percent in women 85 and older, suggesting that the burden may reach an upper limit beyond which the mutation becomes incompatible with life. Earlier cytogenetic work using fluorescence in situ hybridization found LOX in roughly 1.7 percent of lymphocytes from young women but in approximately 22 percent of centenarian women, underscoring the age gradient across the entire human lifespan.

LOX is not a random event. Female cells carry two X chromosomes that differ fundamentally in their epigenetic state: one remains transcriptionally active, while the other is condensed into the inactive X, or Barr body, silenced by the long non-coding RNA XIST during early embryogenesis. Mathematical modeling of LOX profiles from UK Biobank participants has revealed that losing one X chromosome is far more plausible than losing the other, with two distinct LOX events occurring at rates varying by a factor of 140. Three non-exclusive models have been proposed to explain this bias. The first holds that the inactive X is preferentially lost, because retaining the active X maintains adequate gene dosage for cell survival. The second posits that one parentally derived X—either maternal or paternal—is preferentially affected regardless of its activation status. The third, building on both, suggests that the inactive X derived from one specific parent is the most vulnerable target. Single-cell analyses have recently provided evidence that the inactive X is always the chromosome affected by LOX, though the population-genetic modeling paradoxically predicts that the rate of active X loss is more than 70-fold higher than the rate of inactive X loss per year, indicating that both states are substantially impacted at the mutation stage.

What makes the X chromosome, particularly the inactive copy, so prone to vanishing? Cell division is the primary source of chromosome loss, and cytogenetic analysis of cultured lymphocytes from a healthy 62-year-old woman found that the X chromosome represented one-quarter of all lagging chromosomes—six times more often than would be expected by chance. The inactive X carries several intrinsic liabilities. It preferentially undergoes premature centromere division, the precocious loss of centromeric cohesion that drives nondisjunction, likely because XIST RNA actively repels cohesin complexes from the chromosome. It localizes to the nuclear periphery, where chromosomes take longer to align at metaphase and are more likely to be missegregated and incorporated into micronuclei. It replicates late in S phase and accumulates a disproportionately higher mutational load, driven in part by preferential de novo LINE-1 retrotransposon insertions originating from the active X and autosomes. DNA repair is also less efficient in the inactive X than in the active X, and the structural and epigenetic stability of the active X is rigorously safeguarded by master factors such as the deacetylase SIRT7.

A particularly elegant recent discovery has added a structural dimension to this vulnerability. The inactive X is partitioned into two extremely large intervals called superdomains, separated by a boundary containing the microsatellite repeat DXZ4, and it carries superloops—large chromatin loops anchored at DXZ4 that also interact with two other X-linked microsatellite repeats, FIRRE and ICCE. Copy number variations of ICCE and DXZ4 have been found to strongly associate with LOX levels across both UK Biobank and All of Us cohorts. Women with longer arrays of ICCE or DXZ4 show increased LOX in cis, and the effects of these two repeats are additive, contributing to a fourfold range in the likelihood that different X chromosomes will be lost. Two variants within the DXZ4 array that appear to reduce CTCF and YY1 binding disrupt the effect of DXZ4 on LOX, suggesting that increased CTCF and YY1 binding at longer DXZ4 arrays reduces the segregation fidelity of the inactive X. While this hub of repeats is dispensable for the establishment of X chromosome inactivation, it appears essential for the mitotic stability of the silenced chromosome.

Beyond mutation, LOX clones must expand to become clinically relevant, and this expansion is governed by a competitive landscape that researchers have begun to describe as X chromosome-linked competition. The maternal and paternal X chromosomes in any given woman differ by at least 150 non-synonymous coding variants and thousands of non-coding variants, creating fitness differences between clones that retain different parental copies. Three mutually non-exclusive competition scenarios operate in LOX carriers: competition between 45,X cells retaining either the maternal or paternal X, competition between 45,X cells retaining either the active or inactive X, and competition between 45,X and normal 46,XX cells. Evidence from allelic shift analyses has shown that 44 independent X-linked alleles and 59 coding variants exhibit significant shifts in LOX individuals, suggesting they confer differential selective advantages. Simulation work has revealed that loss of the inactive X drives a twofold higher fitness effect compared with loss of the active X, with growth rates of 15.7 percent and 7.3 percent per year respectively, and longitudinal data from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening study indicate that the average population expansion rate of LOX is 0.3 percent per year of age.

The clinical associations of LOX are diverse and increasingly well documented. Detectable LOX is separately associated with increased risk of leukemia overall, acute myeloid leukemia, and chronic lymphoid leukemia by 1.7-fold, 1.9-fold, and 3.3-fold respectively, and expanded LOX confers even greater risks—6.3-fold for leukemia overall, 10.6-fold for acute myeloid leukemia, and 14.7-fold for chronic lymphoid leukemia. These associations with expanded LOX and leukemia risk in UK Biobank persisted even after removing individuals with other clonal hematopoiesis events, indicating that LOX is independently associated with leukemia risk. Beyond hematological malignancy, expanded LOX is associated with heightened risk of systemic sclerosis in Japanese cohorts and with roughly threefold increased risk of Graves’ disease and rheumatoid arthritis in elderly participants of cross-biobank studies. LOX is also associated with increased risk of bacterial pneumonia, a 3.9-fold increase in arterial embolism and thrombosis in the Million Veteran Program, and significantly increased rates of suicide in a large retrospective Japanese study. Notably, women with LOX present in at least 0.87 percent of cells have a 2.16-fold lower likelihood of natural conception, suggesting that even non-expanded LOX may carry pathogenic weight.

The functional consequences of losing an entire chromosome are profound and operate through multiple interconnected mechanisms. LOX entails deletion of hundreds of dosage-sensitive genes, including those in the pseudoautosomal regions and X-Y gametologs that are normally expressed from both copies in both sexes. In response, mammalian stem cells upregulate the remaining active X through a process called X chromosome upregulation, compensating for the dosage imbalance, though this compensation is clearly insufficient to fully rescue the monosomic state. More surprisingly, the historically dismissed inactive X is now recognized as a functionally active chromosome that modulates the expression of autosomal genes in trans. Among four cell types of hematopoietic lineage, the inactive X elicits cell-type-specific responses on autosomal and active X gene expression, and 45,X human induced pluripotent stem cells derived from 46,XX fibroblasts show dysregulated expression of several autosomal genes with regulatory functions. The inactive X also serves as a female-specific buffering mechanism against recessive mutations and as a reservoir dynamically used along neural differentiation, and it tends to become partially reactivated with aging, overexpressing immunity-related genes that increase susceptibility to autoimmunity while simultaneously counteracting age-related cognitive decline in mouse models.

Perhaps the most tantalizing question is whether LOX is a cause or a consequence of biological aging. Preliminary results have shown that subjects with detectable LOX display a 1.82-fold acceleration in DNA methylation-based epigenetic clocks, suggesting LOX is a driver of epigenetic aging. One-directional Mendelian randomization did not suggest a causal relationship between telomere attrition and LOX; instead, telomere length reduction in LOX individuals may be a marker of past clonal growth and expansion. As telomere erosion is itself a driver of aging, reduced telomere length in LOX clones could lead to further genomic instability and the acquisition of additional aging hallmarks. Together, these findings point toward a self-reinforcing loop in which aging promotes the acquisition and expansion of LOX through mechanisms such as inflammaging and declining p53 function, while LOX in turn accelerates biological aging through telomere shortening and epigenetic clock acceleration. Whether modulating LOX burden can be exploited to prevent or mitigate disease in women remains an open question, but the field is moving rapidly from observational studies toward a data-rich discipline with the potential to reshape our understanding of sex differences in aging and age-related disease.

Subject of Research: Loss of the X chromosome in aging women's blood cells as an exemplar of genome aging

Article Title: Genome Aging in Women: Hematopoietic Loss of X Chromosome as An Exemplar

Article References: Guo, X., & Dai, X. (2026). Genome Aging in Women: Hematopoietic Loss of X Chromosome as An Exemplar. Aging Cell, 25(11), Article e70778. https://doi.org/10.1111/acel.70778

Image Credits: AI Generated

DOI: 10.1111/acel.70778

Keywords: loss of X chromosome, genome aging, clonal hematopoiesis, mosaic chromosomal alterations, inactive X chromosome, X chromosome inactivation, cell competition, leukemia risk, autoimmunity, aging women, UK Biobank, somatic mutation