Random X-Chromosome Inactivation in Hemophilia A Carriers Does Not Fully Attenuate Bleeding Risk

Background: In contrast to males with the X-linked bleeding disorder Hemophilia A (HA), bleeding in female carriers is attenuated by the presence of a wild type F8 allele. While low FVIII activity (FVIII:C) is more predictive of bleeding tendency, hemophilia carriers with "normal" FVIII:C can have bleeding sequelae. FVIII:C in HA carriers is influenced by X-chromosome inactivation (XCI), a process that silences one X in XX females. XCI results in cells that have randomly inactivated either parental X, although skewed XCI patterns occur commonly in ~25% of the general population. How frequently XCI skewing in HA carriers leads to preferential activation of the X carrying the F8 mutation, increasing bleeding tendency, is poorly understood. We sought to determine whether XCI skewing further differentiates bleeding risk in HA carriers.

Methods: Patients were enrolled in this prospective, observational cross-sectional study with informed consent when presenting for F8 genotype analysis through the My Life, Our Future (MLOF) study. Standardized Condensed MCMDM-1 and Pediatric Bleeding Questionnaire (PBQ) bleeding assessment tools (BAT) and Pictorial Blood-loss Assessment Charts (PBAC) were completed at the time of the visit. Hemophilia carrier status was confirmed by documented F8 mutation. One-stage FVIII:C assays and chromogenic FVIII assays were performed on all subjects. XCI skewing was assessed using established assays at the Androgen Receptor (AR) and Fragile X Mental Retardation (FMR1) loci.

Results: 98 individuals have been evaluated to date. 65 (46 adults, 19 adolescents and children) are confirmed F8 mutation carriers, while 33 (17 adults) are confirmed non-carriers. 39 (30 adults) have F8 mutations associated with severe HA, and 26 (16 adults) have non-severe F8 mutations. 21/46 (46%) adult HA carriers have BAT scores >4. BAT scores are significantly higher (5 vs. 1.8, p=0.0452) in adult HA carriers compared to age-matched controls (39.9 vs. 35.5, p=0.24). PBAC and BAT correlate well in adult F8 carriers (r =0.62, p<0.0001). Among adolescents, BAT scores are similar between controls and carriers; none of the HA carriers have BAT scores >2.

FVIII assessment by one-stage and chromogenic FVIII assays are highly correlated (p<0.0001), with the notable exception of one carrier whose assays are markedly discrepant (19% chromogenic vs 108% one-stage). Her F8 mutation (p.Glu1988Argfs*4), is associated with severe HA and maps within the A3 domain. Although mutations in this domain can result in discrepant one-stage/chromogenic activity, such an association has not been previously reported for this specific mutation. As chromogenic results more accurately reflect bleeding tendency in this carrier, utilization of both activity assays will capture more individuals at potential risk.

Notably, only one third (7/21) of adult HA carriers with significant bleeding have FVIII:C below 40%. 4/21 have FVIII:C in the 40-60% range, and 10/21 have "normal" levels >60%, indicating that FVIII:C levels alone cannot predict bleeding risk in all carriers. Indeed, for all adult HA carriers (n=46), chromogenic and one-stage FVIII assays correlate only weakly with BAT (r=-0.29, p=0.053) and PBAC (r=-0.27, p=0.072) scores. However, the correlation for severe F8 mutation carriers (n=30) is stronger between BAT scores and chromogenic assays (r=-0.37, p=0.0464), and PBAC scores and one-stage assays (r=-0.36, p=0.0493).

XCI assays are informative in 44 of 46 adult HA carriers. 21/44 adult carriers have skewed XCI (>70:30) and 9 are extremely XCI skewed (>80:20). This level of XCI skewing exceeds that in the general population. Although random XCI is frequently proposed to attenuate carrier bleeding phenotype, 9/23 HA carriers with random XCI have BAT scores >4, including 2 with non-severe F8 mutations predicted to have a mild phenotype. Adolescent HA carriers have similar proportions of XCI skewing and similar FVIII activity to adult HA carriers, yet their reported bleeding is indistinguishable from age matched controls.

Conclusion: Contrary to current dogma, bleeding tendency in HA carriers is not rare, and random X-chromosome inactivation is not fully protective. Correlation of bleeding scores and FVIII:C in adult HA carriers is strengthened when stratified by F8 mutation severity. Additional factors must influence hemostasis for HA carriers with random XCI.