Tissue consistent correlation of factor VIII activity levels with the pattern of X-chromosome inactivation in female carriers of a severe pathogenic variant

BETIANA ZIEGLER; VANINA MARCHIONE; CARLOS DE BRASI; LILIANA ROSSETTI; DANIELA NEME; PAMELA RADIC; MARTÍN ABELLEYRO; ENRIQUE MEDINA ACOSTA

Congreso; WFH2022 World congress; 2021

Introduction: In female carriers (FC), theskewing of X-chromosome inactivation (XCI) against the normal F8 allele is themost typical cause of the haemophilia phenotype. The FVIII:C level inverselycorrelates with the XCI status in symptomatic FC. In a prior study, we deviseda Vshaped correlation model that incorporates the variance in FVIII:C levels innon-carriers. Here, we implemented a biphasic V-model to assess the strength ofthe relationship between the FVIII:C levels and the patterns of XCI in a cohortof FC for whom the haplotype phase of the causative variant is unknown.Methods: The XCI status was investigated in DNA from peripheral blood (PB) from91 FC, 36 non-carriers, and oral mucosa (OM) from 17 FC by genotyping the XCIsurrogate markers at the AR (Xq) and RP2 (Xq) genes. Results: The patterns ofXCI reported by the AR and RP2 surrogate markers were highly concordant ineither PB (n = 84, Spearman r = .7843 [CI95:.6817–.8567], p < .0001) or OM(n = 22, Spearman r = .8238 [CI95:.6080–.9263], p < .0001). The observedcombined heterozygosity for the two markers achieved 100% (n = 155). In PB, weobserved no difference (Mann-Whitney-test p > .05) in the FVII:C/XCIcorrelation using the V-model in FC (Observed-Expected differences [O-E]:11.48, n = 91) versus the mean-model in non-carriers (O-E:10.25, n = 36).We expectthis because in non-carriers the skewing of XCI does not affect F8 expression.However, significant differences were observed from the V-model in FC asweighted against the meanmodel in FC (O-E:22.17, p < .004) or the A-model inFC (without biological basis) (O-E:18.06, p < .0007). In OM, we also see nodifference from the V-model in FC (O-E:8.58, n = 17) versus the mean-model innon-carriers. The difference was significant when weighed against themean-model in FC (O-E:15.94, p < .03) or the A-model in FC (O-E:24.24, p< .0095). Conclusions: Because XCI skewing may act for-or-against the normalF8-allele in heterozygous carriers, we planned a V-shaped FVIII:C/XCIcorrelation model centred on (FVIII:C- carrier- mean; XIP:50%) in which eachline of V (descending/ascending) represents an X-chromosome haplotype phase.The V-model accurately reports the correlation between the whole range ofFVIII:C levels and XCI’s extent in FC, as tested in two tissue types.