More than 2000 HEMA pathogenic variants have been described in the 186 kb F8 causative gene, including inversions, deletions, insertions, duplications, and a wide range of point mutations [4]. Prenatal and preimplantation genetic testing remain crucial enabling tools for at-risk families to avoid affected pregnancies. Whenever possible, PGT-M for HEMA should include direct pathogenic variant detection together with linkage analysis of markers flanking the pathogenic variant, to avoid misdiagnosis due to ADO at the pathogenic variant site. Approximately 45% of severe HEMA cases are caused by Inv22, a ~ 575 kb paracentric inversion within chromosome Xq28 that truncates the F8 gene, mediated by non-allelic homologous recombination between duplicons within intron 22 and distally towards the telomere [1]. Chen et al. successfully demonstrated detection of the Inv22 pathogenic variant in 15 whole genome amplified single lymphocytes of an Inv22 carrier cell line using LD-PCR [10]. When the LD-PCR assay was applied to an actual PGT-M case, complemented with linkage analysis using four informative markers, ADO of the pathogenic variant was detected in one of eight blastomeres. If not for the concurrent analysis of linked markers, this carrier embryo would have been misdiagnosed as normal.

Given the large size of the Inv22 inversion, it is important to ensure that sufficient markers are available within and flanking the inversion breakpoints to ensure that the linkage-based PGT-M assay is robust to amplification failure, allele dropout, and/or unobserved marker-variant recombination. Bui et al. reported using four linked polymorphic STR markers (comprising two telomeric, one intragenic, and one centromeric markers) in 12 linkage-based PGT-M cases for HEMA in Vietnam [12]. However, not all markers were found to be informative in every couple, resulting in some cases where no informative marker was available on one flank of the pathogenic variant, running the risk of misdiagnosis due to unobserved recombination between pathogenic variant and marker on the other flank. In this study, we demonstrate PGT-M of this pathogenic variant hotspot using a 13-marker panel that straddles the almost 600 kb inversion interval. This STR panel, which is highly polymorphic in the Caucasian and Chinese populations [5], has now been shown to be highly polymorphic in a third population group, the majority Kinh ethnic group in Vietnam. In fact, these markers displayed generally higher observed and expected heterozygosity values in the Kinh compared to the Chinese and Caucasian populations (Table 1) and can be used in most if not all at-risk couples regardless of pathogenic F8 variant.

Although intragenic markers have been recommended as the only option for marker choice for PGT-M of HEMA [13], extragenic markers are suitable alternatives when all intragenic markers are uninformative, so long as they are tightly linked and < 1 Mb away from the pathogenic variant to minimize the likelihood of marker-variant recombination to ≤ 1%. In the case of the Inv22 pathogenic variant, however, informative F8 intragenic markers alone may be insufficient to exclude meiotic recombination occurring within the ~ 575 kb inversion interval. The highly polymorphic 13-marker panel increases the likelihood of finding informative markers within as well as flanking either end of the inversion, thus maximizing ability to detect any recombination occurring within the inversion interval. Importantly, despite straddling the hotspot inversion’s duplicons, the distance between the panel’s two most distant markers (HEMA154498.9 and DXS1073) is only ~ 1 Mb (Fig. 1), minimizing the likelihood of recombination between any intra-inversion site and flanking marker to ≤ 1%.

In the first HEMA PGT-M case conducted using the single-tube 13-marker PCR panel as a standalone assay to detect the pathogenic F8 Inv22 variant, four fully informative markers were identified in the couple. They included two markers which were within the Inv22 interval (F8Int21, F8Int13.2), one telomeric (HEMA154498.9), and one centromeric (REN90682). In addition, five markers were partially informative in the couple, of which two were within the inversion interval (F8Int1, HEMA154130.5) and three centromeric (F8Int22, F8Int25.2, REN90833). Trophectoderm samples from three day-5 embryos were analyzed using these nine markers, and all embryos were unequivocally diagnosed to be unaffected, each having inherited the non-recombinant maternal low-risk allele.

In conclusion, sufficient tightly linked informative markers can be reliably used in PGT-M of HEMA, either as a standalone linkage-based test or to complement pathogenic F8 variant detection.