Journal of Thrombosis and Haemostasis, 8: 783–789

DOI: 10.1111/j.1538-7836.2010.03768.x

ORIGINAL ARTICLE

Preimplantation genetic diagnosis for hemophilia A using indirect linkage analysis and direct genotyping approaches A. D. LAURIE,* A. M. HILL,  J. R. HARRAWAY,* A. P. FELLOWES,* G. T. PHILLIPSON,  P. S. BENNY,  M . P . S M I T H * and P . M . G E O R G E * *Canterbury Health Laboratories, Christchurch; and  Repromed Christchurch, Hiatt Chambers, Christchurch, New Zealand

To cite this article: Laurie AD, Hill AM, Harraway JR, Fellowes AP, Phillipson GT, Benny PS, Smith MP, George PM. Preimplantation genetic diagnosis for hemophilia A using indirect linkage analysis and direct genotyping approaches. J Thromb Haemost 2010; 8: 783–9.

Summary. Background: Preimplantation genetic diagnosis (PGD) is an appealing option for couples at risk of having a child with hemophilia A (HA). Although many clinics offer PGD for HA by gender selection, an approach that detects the presence of the underlying F8 mutation has several advantages. Objectives: To develop and validate analysis protocols combining indirect and direct methods for identifying F8 mutations in single cells, and to apply these protocols clinically for PGD. Methods: A panel of microsatellite markers in linkage disequilibrium with F8 were validated for single-cell multiplex polymerase chain reaction. For point mutations, a primer extension genotyping assay was included in the multiplex. Amplification efficiency was evaluated using buccal cells and blastomeres. Four clinical PGD analyses were performed, for two families. Results: Across all validation experiments and the clinical PGD cases, approximately 80% of cells were successfully genotyped. Following one of the PGD cycles, healthy twins were born to a woman who carries the F8 intron 22 inversion. The PGD analysis for the other family was complicated by possible germline mosaicism associated with a de novo F8 mutation, and no pregnancy was achieved. Conclusions: PGD for the F8 intron 22 inversion using microsatellite linkage analysis was validated by the birth of healthy twins to one of the couples. The other familyÕs situation highlighted the complexities associated with de novo mutations, and possible germline mosaicism. As many cases of HA result from de novo mutations, these factors must be considered when assessing the reproductive options for such families. Keywords: F8, germline, microsatellite, mosaicism, mutation, PGD.

Correspondence: Andrew D. Laurie, Molecular Pathology, Canterbury Health Laboratories, PO Box 151, Christchurch, New Zealand. Tel.: +64 3 3786001; fax: +64 3 3640545. E-mail: [email protected] Received 13 December 2009, accepted 14 January 2010  2010 International Society on Thrombosis and Haemostasis

Introduction Hemophilia A (HA) is an X-linked recessive bleeding disorder caused by mutations in the factor VIII gene (F8) [1]. Factor VIII is an essential component of the clotting cascade, where it acts as a cofactor for the serine protease FIX, forming the tenase complex on activated platelet membranes. The HA phenotype can be classified as mild, moderate, or severe, depending on the level of FVIII activity detected in plasma. Although > 1000 unique HA-causing F8 mutations have been reported, a large proportion (40–45%) of severe HA cases result from an inversion affecting intron 22 of F8 [2,3]. The high morbidity associated with severe HA has made prenatal diagnosis, and more recently preimplantation genetic diagnosis (PGD), an appealing option for carrier females and their partners who want to avoid having an affected child [4,5]. Many clinics offer PGD for HA, using a strategy applied to a range of X-linked recessive disorders; that is, gender selection by fluorescence in situ hybridization, and transfer of only female embryos [6]. However, an analysis that allows the diagnosis of the HA status of every embryo (i.e. affected male, carrier female, or normal/not affected) is more desirable, because the number of embryos available for transfer from each cycle is increased (75%, as compared with 50% for females only). In addition, couples prefer to know that they have the possibility of conceiving a healthy child of either gender. Several laboratories have reported successful PGD cycles in which embryos were screened for a specific familial F8 mutation. The first live birth following mutation-specific PGD for HA was reported in 2006, to a woman carrying the F8 R1689C mutation [7]. Several other reports have described PGD for HA targeting a range of F8 mutations, yet, to our knowledge, a live birth following PGD for the common intron 22 inversion has not yet been reported [8–11]. GutierrezMateo et al. [12] reported a referral for an F8 intron 22 inversion PGD, but, at the time of publication, no cycles had been performed. The F8 intron 22 inversion mutation cannot be directly detected by single-cell polymerase chain reaction (PCR),

784 A. D. Laurie et al

because the long-distance PCR (LD-PCR) assay requires 300 ng of DNA template, whereas a single blastomere has only 6 pg of DNA [13]. Although a novel method for detecting the inversion using inverse shifting PCR (IS-PCR) was recently described, this approach requires 2 lg of sample DNA and is also not suitable for single-cell analysis [14]. Alternatively, a whole genome amplification (WGA) step could be combined with direct detection of the inversion by either LD-PCR or ISPCR, but the high rate of amplification bias inherent in WGA, estimated to be  30%, means that this is not a viable option [15]. Therefore, a linkage approach using microsatellite markers located near the F8 gene is the preferred method for identification of the F8 allele carrying the intron 22 inversion mutation. Several variable-length short tandem repeats have been described that are located within or flanking the F8 gene and would be suitable for linkage analysis [11,16,17]. A panel of three or four informative markers can be amplified together in a single multiplex PCR from each lysed blastomere, and the derived haplotype is used to infer the presence or absence of the F8 mutation. The use of several markers provides redundancy in the analysis, so that the status of a sample may be determined even if some of the markers fail to amplify or are subject to PCR amplification bias resulting in allele dropout (ADO), a known problem associated with single-cell analysis [18]. Here we report the results of cleavage-stage PGD for HA using a multiplexed panel of F8-linked microsatellite markers, which can be combined with direct mutation genotyping if appropriate. We describe our experience with two families. In one family, we report the first live births to a woman carrying the F8 intron 22 inversion following PGD by linkage analysis. In the second family, we describe the complexities associated with PGD for a de novo F8 mutation with possible germline mosaicism. Materials and methods In vitro fertilization (IVF) and embryo biopsy

Fertilization was achieved by intracytoplasmic sperm injection, to minimize the chances of contamination by spermatozoa embedded in the zona pellucida at the time of blastomere aspiration. Embryos were cultured in cleavage culture medium (Cook, Sydney, NSW, Australia) up to day 3 postinsemination (p.i.). Embryo biopsy was performed on the morning of day 3 p.i. in Cook Ca2+/Mg2+-free Biopsy Medium. After biopsy, embryos were transferred to Cook Blastocyst Medium and kept in this medium until transfer. Embryo biopsies were performed using a Zilos-tk Zona Infra-red Laser Optical System (Hamilton Thorne, Beverly, MA, USA) for piercing of the zona pellucida. The biopsy pipette was inserted through the hole, and a blastomere was removed by aspiration. Usually, two cells were removed from embryos with seven blastomeres or more and only one cell from the others, unless the first cell removed had lysed or had no visible nucleus. Blastomeres were rinsed in a drop of

phosphate-buffered saline supplemented with 0.1% poly(vinyl alcohol) (Medicult, Jyllinge, Denmark), using a mouth-controlled, finely pulled glass pipette, before being transferred into a 0.5-mL microcentrifuge tube containing 9 lL of lysis buffer [5 lM sodium dodecylsulfate, 100 lg mL)1 proteinase K (Qiagen, Hilden, Germany), and 10 lM Tris, pH 8.4]. All blastomeres were placed in separate tubes and analyzed individually. For each blastomere, a small volume of rinsing medium was transferred similarly and used as a negative PCR control. The tubes containing lysis buffer and cells were kept on a refrigerated rack at 4 C during the biopsy procedure, incubated at 50 C for 30 min for cell lysis, and then heated at 95 C for 10 min to inactivate the proteinase K. Samples were then transported to the molecular laboratory for the PCR and analysis steps. Before biopsy, an assessment of each embryoÕs morphology was assigned using the following criteria: cell number, shape and size of the blastomeres, and degree of extracellular fragmentation. Unaffected embryos with the best morphology were selected for transfer irrespective of their sex or carrier status. Transfers were carried out on either day 4 p.i. or day 5 p.i. Signed informed consent was obtained from all patients, including those whose abnormally fertilized embryos were used for validation of PCR from single blastomeres. The patient information and consent process covers the PGD procedure, including service audit and anonymous data reporting. Multiplex PCR from single cells

To each lysed cell sample (or negative control) was added 20 lL of bulk multiplex PCR mix, so that the final reaction contained 1 · Multiplex PCR Master Mix (Qiagen), 1 · Qsolution (Qiagen), and 200 nM each primer. Primers were supplied by Invitrogen (Carlsbad, CA, USA); sequences of primers are given in Table 1. Thermal cycling was performed on an Eppendorf Mastercycler Gradient (Eppendorf, Hamburg, Germany), with the following parameters: 95 C for 15 min; 10 cycles of 98 C for 1 min, 58 C for 3 min, and 72 C for 1 min; 40 cycles of 94 C for 45 s, 58 C for 3 min, and 72 C for 1 min; and 60 C for 45 min. The FAM-labeled products were resolved on an ABI3130 analyzer (Applied Biosystems, Foster City, CA, USA) by mixing 1 lL of the PCR reaction with 9 lL of Hi-Di formamide containing GS500LIZ size standard (Applied Biosystems), and analyzed using GENEMAPPER 4.0 software (Applied Biosystems). Direct genotyping analysis

The F8:c.6901-2A>G mutation was genotyped directly using a SNaPshot primer extension assay (Applied Biosystems). Primers flanking exon 26 of F8 were included in the multiplex PCR at 200 nM (for family 2 analysis only). Following amplification, 20 lL of each PCR product was purified using a MultiScreen spin plate (Millipore, Billerica, MA, USA), and diluted 20-fold, and 1 lL was then used as the template with F8_6901-2F  2010 International Society on Thrombosis and Haemostasis

PGD for hemophilia A 785 Table 1 Sequences of primers used in this work. Product size (bp)

Primer*

Sequence

F8C-IVS13F F8C-IVS13R -FAM F8C-IVS22F F8C-IVS22R -FAM DXS1073F DXS1073R -FAM F8ex26F F8ex26R F8_6901-2F

TGCATTCAACTGTACATAATGTATCTT 153 CCAAATTACAGATTGAATAAGCCTAG ATTAATGCCCACATTATAGACTCTC AATAAGACCCTTAGCTGTTTCAT

206

GCCCTCTCCGAGTTATTAC ATTGGTGGCCTTTGAAACA

128

CCAATAAATGCTATCTTTCCT TGACGGCAGTGGCAGGTGCT CAATAAATGCTATCTTTCCTCTTTC

217

*FAM indicates that the primer is labeled with a FAM fluorophore on the 5¢-end.

primer (at 200 nM) in a 10-lL reaction using the SNaPshot reagents as per the kit protocol. Thermal cycling conditions were as follows: after 25 cycles of 96 C for 10 s, 50 C for 5 s, and 60 C for 30 s, 1 lL of shrimp alkaline phosphatase (USB, Cleveland, OH, USA) was added to each reaction and incubated at 37 C for 60 min, and then at 75 C for 15 min. Reactions were resolved on the ABI3130 as described above, and analyzed using GENEMAPPER software. Mutation analysis for F8 intron 22 inversion

Screening of patient DNA samples for the F8 intron 22 inversion mutation was performed according to a modification of the LD-PCR protocol of Liu et al. [13,19]. Results Validation of multiplex PCR from single cells

Five polymorphic microsatellite markers located within or flanking F8 were assessed for linkage analysis for HA PGD [16]. The two intragenic markers (F8C-IVS13 and F8C-IVS22), together with DXS1073, located 235 kb from F8 (Fig. 1), were informative for the two families described here. Therefore, these markers were optimized for single-cell multiplex PCR. The cell lysis and amplification protocol for single-cell analysis was validated by assessing amplification from 130

100 kb (distal)

(proximal)

F8

DXS1073

F8C-IVS13 F8C-IVS22

Fig. 1. Schematic diagram showing location of the microsatellite markers within or near the F8 gene that were used in the linkage analysis for preimplantation genetic diagnosis for hemophilia A. The DXS1073 marker is located 235 kb on the centromeric (proximal) side of F8.  2010 International Society on Thrombosis and Haemostasis

single buccal cells (from the same individual), and 161 blastomeres derived from abnormally fertilized embryos from patients attending the fertility clinic, analyzed in 18 separate experiments. Data for each marker are shown in Table 2, indicating an amplification success of 72–82%, with an ADO rate of 7–13%. Successful amplification of two markers is sufficient for a diagnosis, and on this basis we could assign status to approximately 80% of all cells analyzed; this was deemed to be acceptable for clinical utility. IVF statistics

Family 1 had three cycles of IVF/PGD in 2006–2007 and achieved a dizygotic pregnancy on their third cycle. Family 2 had a single IVF/PGD cycle in 2007 but did not achieve a pregnancy. Details of these four cycles are shown in Table 3; for the 32 blastomeres analyzed in these cycles, the amplification success across all markers was 63–88%, with an ADO rate of 0–16%. A diagnostic result was obtained for 84% (27/32) of blastomeres analyzed, allowing diagnosis of the mutation status of 77% (17/22) of embryos. Analysis details – family 1

The female partner in this family (III:2 in Fig. 2) carries the F8 intron 22 inversion. The mutation was inherited from her father (II:1), who had severe HA and is now deceased. As this couple had no children, and no material from the deceased father was available, the phasing of the F8 markers with the intron 22 inversion was determined by analyzing II:3, who is the cousin of II:1, who also has HA and the intron 22 inversion (FVIII:C < 1 IU dL)1), along with his daughter III:3, who is a carrier (Fig. 2). The DXS1073 and F8C-IVS13 markers were fully informative for this couple, and the F8C-IVS22 marker was partially informative but had a unique product size for the allele in cis with the mutation (Fig. 2). PGD analysis for this couple focused on detecting the DXS1073 and F8C-IVS13 normal maternal alleles (135 and 153), as the presence of these products would identify a normal male or female. If the affected maternal allele was detected in an embryo, there had to be the clear presence of at least two products from the paternal allele for the embryo to be considered for transfer (as a carrier female). Across the three cycles, we obtained a clear diagnostic result for 15 embryos from a total of 19 that were biopsied (79%). The distribution of the four allele combinations followed approximately Mendelian ratios, with three embryos identified as affected males (20%). Despite two apparently healthy embryos being transferred in each of the first two cycles, no pregnancy was established. After the third cycle, a twin pregnancy was confirmed, and a boy and girl were delivered at almost full term. DNA was extracted from the cord blood of both infants and analyzed directly by LD-PCR specific for the F8 intron 22 inversion and for the panel of markers. This analysis confirmed that neither infant had the mutation, and the haplotypes agreed with the PGD result (Fig. 2).

786 A. D. Laurie et al Table 2 Summary of single-cell multiplex polymerase chain reaction data across 18 separate experiments; all three markers were amplified in a single reaction, as described in Materials and methods.

Microsatellite marker

Cells with successful amplification

Cells in which amplification failed

Cells with allele dropout

Analysis of 130 single buccal cells, no. (%) DXS1073 107 (82) 19 (15) – F8C-IVS13 94 (72) 19 (15) 17 (13) F8C-IVS22 96 (74) 18 (14) 16 (12) Analysis of 161 blastomeres derived from abnormally fertilized embryos, no. (%) DXS1073 116 (72) 31 (19) 14 (9) F8C-IVS13 120 (75) 29 (18) 12 (7) F8C-IVS22 124 (77) 24 (15) 13 (8)

Analysis details – family 2

This couple have a son with severe HA (FVIII:C of 1 IU dL)1) and an identified F8 mutation (c.6901-2A>G) predicted to cause aberrant splicing of the F8 mRNA transcript [20]. There was no family history of HA, and this mutation could not be detected in his mother (I:2 in Fig. 3) in DNA extracted from peripheral lymphocytes. Several scenarios could explain the origin of this de novo mutation, and each would indicate a different risk of recurrence for future children. A negligible risk would be implied if the mutation had arisen in an early postfertilization mitosis in the affected child, but as this would be expected to result in somatic mosaicism in the boy, which was not evident (in lymphocytes), this scenario is unlikely [21]. Alternatively, the mutation could have occurred during the process of gametogenesis in his mother (I:2), resulting in germline mosacism, which would be associated with a variable risk of recurrence, depending on whether the mutation occurred at an early division (high risk) or a late division (low risk). A third possibility is that I:2 is a somatic plus germline mosaic for the c.6901-2A>G mutation, because it arose very early in her own embryogenesis, before separation of the somatic and germ cell lineages [21,22]. Such a scenario could result in a significant risk of recurrence of HA in future children.

Analysis of the F8 microsatellite panel for this couple and their affected son allowed phasing of the product sizes to identify the normal F8 alleles and the ÔaffectedÕ/at-risk allele associated with the c.6901-2A>G splicing variant in the son (Fig. 3). Although these data would be sufficient for performance of PGD for this couple by excluding males with the atrisk allele, we were interested in directly genotyping the embryos for the c.6901-2A>G variant, so that we could: (i) investigate the possibility of germline mosaicism in I:2 and thereby provide advice to this couple regarding future reproductive options; and (ii) increase the available number of embryos for transfer by identifying any embryos with the atrisk haplotype but without the splicing mutation. To perform the direct genotyping analysis, we optimized a primer extension protocol that could be included in the PGD analysis along with the microsatellite sizing. From the coupleÕs IVF/PGD cycle, three embryos were obtained, and on the day 3 biopsy one blastomere was removed from each for the PGD analysis. Only two blastomeres were successfully amplified by PCR. This identified both embryos as male, with each having a different copy of the maternal F8 region. The genotyping analysis revealed that the blastomere with the at-risk allele did not have the c.69012A>G mutation (Fig. 3). Both embryos were therefore classed as Ônot affectedÕ, and although both were transferred, no pregnancy was established. Several months later, the couple conceived naturally and the laboratory received a sample for prenatal analysis. The fetus was female, and carried the at-risk maternal F8 allele, but did not have the c.6901-2A>G mutation. A healthy girl was born at full term. In 2009, the couple again conceived naturally, and prenatal analysis identified a female with the same F8 alleles as their previous daughter, and a normal c.6901-2A genotype. Discussion and conclusion The PGD analyses for these two families has demonstrated the utility of linkage analysis and direct genotyping for HA PGD, adding to the data already reported by other groups [7–9,11]. The use of linked markers to indirectly detect the F8 intron 22 inversion for PGD analysis has been validated by the birth of

Table 3 Summary of the four clinical in vitro fertilization/preimplantation genetic diagnosis cases described in this work

No. of cumulus complexes retrieved No. of oocytes injected No. of two pronuclei No. of embyros biopsied No. of blastomeres analyzed No. of blastomeres diagnosed No. of embryos diagnosed No. of transferable embryos No. of embryos transferred Outcome

Family 1 Cycle 1

Family 1 Cycle 2

Family 1 Cycle 3

Family 2 Cycle 1

19 15 14 7 8 6 4 3 2 No pregnancy

6 6 6 4 5 4 4 4 2 No pregnancy

12 9 9 8 16 15 7 5 2 Twins born in May 2008

3 3 3 3 3 2 2 2 2 No pregnancy

 2010 International Society on Thrombosis and Haemostasis

PGD for hemophilia A 787

1

3

2

4

I

2

1

3

4

II DXS1073 143 F8C-IVS13 155 F8C-IVS22 208 IN22 F8 1

2

3

III 129 DXS1073 F8C-IVS13 151 F8C-IVS22 212 normal F8

143 155 208 IN22

1

135 153 212 normal

DXS1073 143 F8C-IVS13 155 F8C-IVS22 208 IN22 F8

129 161 210 normal

2

IV DXS1073 129 F8C-IVS13 151 F8C-IVS22 212 normal F8

135 135 153 153 212 212 normal normal

Fig. 2. Family 1 pedigree showing product sizes (in bp) for the three F8-linked microsatellite markers used in the hemophilia A (HA) preimplantation genetic diagnosis analysis, and the phasing with the intron 22 inversion mutation (IN22). Solid gray symbols indicate individuals with HA; symbols with solid gray centers indicate female carriers of the intron 22 inversion.

healthy twins in family 1, and direct LD-PCR analysis on DNA derived from cord blood confirmed absence of the inversion mutation in both infants. Significantly, one of the twins is male, an outcome that is not possible with the gender selection method practised by many clinics for HA PGD. Although the PGD analysis for family 2 did not result in a pregnancy from the IVF cycle, the genotype of the two embryos from which we obtained data allowed a reassessment of the risk of this couple conceiving an affected male. In cases of de novo HA such as this, where a new patient has an identified F8 mutation but his mother does not have the mutation, there are three scenarios that can explain the origin of that mutation (as described above), and each scenario has different implications in terms of risk of recurrence for future children [21,23]. Of these, mutations in germline cells, resulting in germline mosaicism, is recognized as the most frequent cause of de novo hemophilia, and was presumed to underlie the origin of the F8 mutation in this case [21]. Germline mosaicism arises from mitotic errors during gametogenesis, and the proportion of gametes affected will depend on at which stage the error occurred. The estimated number of cell divisions undergone by a mature human egg is 23–24, so the fraction of affected eggs will be extremely variable, depending on the point at which the mutation occurred during the process of division [24]. Germline mosaics may also be somatic mosaics if the replication error occurred during the epiblast stage of embryogenesis; the epiblast precedes the endoderm and ectoderm lineages of the embryo,  2010 International Society on Thrombosis and Haemostasis

but also retains the capacity to form germ cells [25]. Several studies have demonstrated somatic mosaicism in the mothers or maternal grandparents of sporadic HA probands, either by detection of a fraction of the mutant allele in lymphocyte DNA by sequence analysis, or by using a more sensitive detection method, such as allele-specific PCR [26–28]. One study found that 13% of mothers or maternal grandmothers of sporadic cases were, in fact, somatic mosaics, emphasizing the fact that mosaicism for F8 mutations is not uncommon [27]. We have not analyzed I:2 with such approaches, and so cannot exclude the possibility that she is a somatic mosaic. Clinical geneticists have calculated the recurrence risk for de novo mutations thought to have occurred during oogenesis by assuming that a mutation has an equal chance of arising at each of the 23 or 24 mitotic divisions that occur during the development of a mature human oocyte, deriving a risk of 4– 5% [23]. However, this is an average value, and the actual risk could be as high as 50% if the mutation occurred during the first mitotic division such that all diploid gametes would be affected, resulting in half of all haploid gametes having the mutation. If a replication error arose during the epiblast stage of embryogenesis in the mother, resulting in somatic plus germline mosaicism, the number of affected gametes could also be as high as 50%, that is, the same as if the woman was a true carrier, giving a risk of having an affected child of 25%. When family 2 members were considering their IVF/PGD treatment, we had genotype information from only one female gamete, that is, the F8 allele with the c.6901-2A>G mutation

788 A. D. Laurie et al 1

2

143 155 208 A

135 159 206 A

I DXS1073 F8C-IVS13 F8C-IVS22 F8:c.6901-2

1

2

131 153 212 A

3

II DXS1073 135 F8C-IVS13 159 F8C-IVS22 206 F8:c.6901-2 G

135 159 206 A

143 155 208 A

135 159 206 A

143 155 208 A

PGD analysis (no pregnancy established) embryo 1

DXS1073 F8C-IVS13 159 F8C-IVS22 206 F8:c.6901-2 A

embryo 2

131 153 212 A

Fig. 3. Family 2 pedigree showing product sizes (in bp) for the three F8linked microsatellite markers used in the hemophilia A (HA) preimplantation genetic diagnosis (PGD) analysis, and the genotype of the F8:c.6901-2 position. The soild gray symbol indicates the son with HA who has the F8:c.6901-2A>G splicing mutation. Data from the PGD analysis are shown, with an arrow indicating the chronological order of this analysis relative to the birth dates of the proband and his sisters (that is, the proband is the oldest child).

identified in their affected son, and we had no evidence that the at-risk haplotype could exist without this F8 mutation. The results of the PGD analysis demonstrated that I:2 had gametes that carried the at-risk haplotype in cis with a normal F8 gene, indicating that the recurrence risk was likely to be significantly less than the maximum risk estimates quoted above. This result may have given this couple the confidence to continue with their family without further intervention, although prenatal analyses were offered and performed for both subsequent pregnancies. Taking the data from this coupleÕs children, their current pregnancy and the embryo analyses together, the atrisk haplotype was seen in three cases without the F8 mutation, and in only one instance in cis with the mutation. Although this is not a statistically significant sample, these data point towards a low level of germline mosiacism, probably arising from a mutation late in gametogenesis in the ovaries of this woman. So-called sporadic cases of HA, defined by a lack of any family history of the disease, have been estimated to account for  50% of patients with severe HA (although only 30% with mild–moderate HA) [29]. For sporadic HA caused by the F8 intron 22 inversion, the mutation can usually be traced back to a meiotic error in the sperm of the probandÕs maternal grandfather [30]. Point mutations may arise in both males and females during embryogenesis or gametogenesis, as described above, and commonly occur at CpG sites (although it should

be noted F8:c.6901-2A>G does not occur at a CpG site) [27]. Kasper and Lin [29] found that the mother was a carrier of the mutation in 56 of 66 families with a single, isolated case of HA or hemophilia B. Thus, 10 of 66 (15%) of these HA pedigrees could be complicated by germline or somatic mosaicism. These situations represent a source of uncertainty for hematologists, clinical geneticists, and genetic counselors, who must provide recurrence risk estimates to parents of a child with HA. Unlike autosomal disorders, where semen analysis can be performed to confirm a mutationÕs origin and estimate the proportion of mosaicism if paternal germline mosaicism is suspected, oocyte harvest and analysis for Xlinked conditions such as HA is not practical clinically [31]. Therefore, the coupleÕs embryos and/or children must be analyzed to infer the presence of maternal germline mosaicism. Given the increasing availability of a range of reproductive options for such families, and the requirement for accurate risk prediction and genetic counseling, thorough consideration and investigation of these complex aspects of heredity is essential. The staff of hemophilia clinics have an important role in providing a modern, comprehensive genetic risk analysis of hemophilia pedigrees, and advising carrier females of their reproductive options. Disclosure of Conflict of Interests The authors state that they have no conflict of interest. References 1 Bowen D. Hemophilia A and hemophilia B: molecular insights. Mol Pathol 2002; 55: 127–44. 2 Kemball-Cook G, Tuddenham EG, Wacey AI. The factor VIII structure and mutation resource site: HAMSTeRS version 4. Nucleic Acids Res 1998; 26: 216–19. 3 Lakich D, Kazazian HH Jr, Antonarakis SE, Gitschier J. Inversions disrupting the factor VIII gene are a common cause of severe haemophilia A. Nat Genet 1993; 5: 236–41. 4 Braude P, Pickering S, Flinter F, Ogilvie CM. Preimplantation genetic diagnosis. Nat Rev Genet 2002; 3: 941–53. 5 Lavery S. Preimplantation genetic diagnosis of haemophilia. Br J Haematol 2009; 144: 303–7. 6 Pickering S, Polidoropoulos N, Caller J, Scriven P, Ogilvie CM, Braude P. Strategies and outcomes of the first 100 cycles of preimplantation genetic diagnosis at the GuyÕs and St. ThomasÕ Center. Fertil Steril 2003; 79: 81–90. 7 Michaelides K, Tuddenham EG, Turner C, Lavender B, Lavery SA. Live birth following the first mutation specific pre-implantation genetic diagnosis for haemophilia A. Thromb Haemost 2006; 95: 373–9. 8 Fiorentino F, Biricik A, Nuccitelli A, De Palma R, Kahraman S, Iacobelli M, Trengia V, Caserta D, Bonu MA, Borini A, Baldi M. Strategies and clinical outcome of 250 cycles of preimplantation genetic diagnosis for single gene disorders. Hum Reprod 2006; 21: 670–84. 9 Sa´nchez-Garcı´ a JF, Gallardo D, Navarro J, Ma´rquez C, Gris JM, Sa´nchez MA, Altisent C, Vidal F. A versatile strategy for preimplantation genetic diagnosis of haemophilia A based on F8-gene sequencing. Thromb Haemost 2006; 96: 839–45. 10 Tomi D, Griesinger G, Schultze-Mosgau A, Eckhold J, Scho¨pper B, Al-Hasani S, Diedrich K, Schwinger E. Polar body diagnosis for hemophilia A using multiplex PCR for linked polymorphic markers. J Histochem Cytochem 2005; 53: 277–80.  2010 International Society on Thrombosis and Haemostasis

PGD for hemophilia A 789 11 Gigarel N, Frydman N, Burlet P, Kerbrat V, Steffann J, Frydman R, Munnich A, Ray PF. Single cell co-amplification of polymorphic markers for the indirect preimplantation genetic diagnosis of hemophilia A, X-linked adrenoleukodystrophy, X-linked hydrocephalus and incontinentia pigmenti loci on Xq28. Hum Genet 2004; 114: 298– 305. 12 Gutie´rrez-Mateo C, Sa´nchez-Garcı´ a JF, Fischer J, Tormasi S, Cohen J, Munne´ S, Wells D. Preimplantation genetic diagnosis of single-gene disorders: experience with more than 200 cycles conducted by a reference laboratory in the United States. Fertil Steril 2009; 92: 1544–56. 13 Liu Q, Nozari G, Sommer SS. Single-tube polymerase chain reaction for rapid diagnosis of the inversion hotspot of mutation in hemophilia A. Blood 1998; 92: 1458–9. 14 Rossetti LC, Radic CP, Larripa IB, De Brasi CD. Developing a new generation of tests for genotyping hemophilia-causative rearrangements involving int22h and int1h hotspots in the factor VIII gene. J Thromb Haemost 2008; 6: 830–6. 15 Renwick PJ, Lewis CM, Abbs S, Ogilvie CM. Determination of the genetic status of cleavage-stage human embryos by microsatellite marker analysis following multiple displacement amplification. Prenat Diagn 2007; 27: 206–15. 16 Harraway JR, Smith MP, George PM. A highly informative, multiplexed assay for the indirect detection of hemophilia A using fivelinked microsatellites. J Thromb Haemost 2006; 4: 587–90. 17 Sa´nchez-Garcı´ a JF, Gallardo D, Ramı´ rez L, Vidal F. Multiplex fluorescent analysis of four short tandem repeats for rapid haemophilia A molecular diagnosis. Thromb Haemost 2005; 94: 1099–103. 18 Ray PF, Handyside AH. Increasing the denaturation temperature during the first cycles of amplification reduces allele dropout from single cells for preimplantation genetic diagnosis. Mol Hum Reprod 1996; 2: 213–18. 19 Kilian NL, Pospisil V, Hanrahan V. Haemophilia A, factor VIII intron 22 inversion screening using subcycling-PCR. Thromb Haemost 2006; 95: 746–7.

 2010 International Society on Thrombosis and Haemostasis

20 Laurie AD, Sheen CR, Hanrahan V, Smith MP, George PM. The molecular aetiology of haemophilia A in a New Zealand patient group. Haemophilia 2007; 13: 420–7. 21 Kasper CK, Buzin CH. Mosaics and haemophilia. Haemophilia 2009; 15: 1181–6. 22 Zlotogora J. Germ line mosaicism. Hum Genet 1998; 102: 381–6. 23 Young ID. Introduction to Risk Calculation in Genetic Counseling, 3rd edn. Oxford: Oxford University Press, 2006. 24 Crow JF. The origins, patterns and implications of human spontaneous mutation. Nat Rev Genet 2000; 1: 40–7. 25 Wylie C. Germ cells. Cell 1999; 96: 165–74. 26 Costa C, Frances AM, Letourneau S, Girodon-Boulandet E, Goossens M. Mosaicism in men in hemophilia: is it exceptional? Impact on genetic counselling. J Thromb Haemost 2009; 7: 367–9. 27 Leuer M, Oldenburg J, Lavergne JM, Ludwig M, Fregin A, Eigel A, Ljung R, Goodeve A, Peake I, Olek K. Somatic mosaicism in hemophilia A: a fairly common event. Am J Hum Genet 2001; 69: 75– 87. 28 Ranjan R, Biswas A, Meena A, Akhter MS, Yadav BK, Ahmed RH, Saxena R. Importance of investigating somatic and germline mutations in hemophilia A: a preliminary study from All India Institute of Medical Sciences, India. Clin Chim Acta 2008; 389: 103–8. 29 Kasper CK, Lin JC. Prevalence of sporadic and familial haemophilia. Haemophilia 2007; 13: 90–2. 30 Rossiter JP, Young M, Kimberland ML, Hutter P, Ketterling RP, Gitschier J, Horst J, Morris MA, Schaid DJ, de Moerloose P, Sommer SS, Kazazian HH Jr, Antonarakis SE. Factor VIII gene inversions causing severe hemophilia A originate almost exclusively in male germ cells. Hum Mol Genet 1994; 3: 1035–9. 31 Edghill EL, Gloyn AL, Goriely A, Harries LW, Flanagan SE, Rankin J, Hattersley AT, Ellard S. Origin of de novo KCNJ11 mutations and risk of neonatal diabetes for subsequent siblings. J Clin Endocrinol Metab 2007; 92: 1773–7.