J Clin Pathol: first published as 10.1136/jcp.40.9.971 on 1 September 1987. Downloaded from

J Clin Pathol 1987;40:971-977

Haemophilia A: carrier detection and prenatal diagnosis by linkage analysis using DNA polymorphism

E G D TUDDENHAM,* ELEANOR GOLDMAN,t A McGRAW,* P B A KERNOFFt From the *Haemostatis Research Group, Clinical Research Centre, Harrow, Middlesex, and the tHaemophilia Centre, Royal Free Hospital, London

SUMMARY Restriction fragment length polymorphisms (RFLPs) within or close to the factor VIII locus are very useful for genetic linkage analysis. Such RFLPs allow a mutant allele to be tracked in a family, segregating A even when, as is usually the case, the precise causing failure to synthesise factor VIII is unknown. To date two markers tightly linked to the factor VIII locus have been described, one ofwhich is highly polymorphic and therefore informative in most kindreds. A significant crossover rate, however, does not make diagnosis absolute. Three intragenic RFLPs have been defined, which, taken together, are informative in about 70% of women, providing virtually deterministic genetic diagnosis. copyright. Diagnosis of the carrier state is an urgent clinical biopsy samples were taken by the referring centre and problem in families with a known case ofhaemophilia shipped as whole blood or tissue frozen on dry ice. A. For every affected male, on average eight female relatives will be found who have a greater or lesser SEPARATION OF DNA genetic risk of carrying the mutant allele. Risk assess- Blood anticoagulated with edetic acid in standard ment was based until recently on pedigree analysis haematology sample tubes was frozen and stored at

combined with factor VIII and Von Willebrand fac- -40'C until processed. The choice of anticoagulant http://jcp.bmj.com/ tor antigen assays.' At best, about 85% accuracy of is dictated by the fact that some restriction enzymes, discrimination from the phenotype assays was including XbaI, work poorly or not at all on hepa- obtained. As a result many females at relatively low rinised samples. No deterioration of DNA over time risk opted for definitive antenatal diagnosis by means or after shipment on dry ice from Australia to the of factor VIII bioassay or antigen assay of fetoscopic United Kingdom was noted. DNA was separated as blood sample taken in the second trimester.2 The described45 and stored at -20°C until analysed. introduction of linkage analysis based on the use of intragenic DNA probes and linked DNA markers has DNA PROBES on September 29, 2021 by guest. Protected greatly improved definitive diagnosis.3 In this paper DX13 is a fragment of the isolated we describe the DNA probes available for diagnosis from an X genomic library6 that has been shown to be of and give examples of their use in linked to haemophilia A7 and to the factor VIII families studied over the past two years. gene.8 The restriction fragment length polymorphism (RFLP) is detected with BglII. StJ4 is a similar fragment derived from the same Material and methods library9 and shown to be linked to haemophilia A"0 and to the factor VIII gene.8 This locus is highly poly- Families with at least one confirmed case of hae- morphic with 10 alleles detectable using TaqI. mophilia A were referred for genetic counselling to Factor VIII exon 17,18 This is a genomic Stu I/Sca the Haemophilia Centre, Royal Free Hospital, by I fragment that detects a BclI RFLP in the adjacent physicians from the United Kingdom and world wide. intron 18.8 Fig 1 shows the location of this poly- Local residents attended in person for counselling morphism. and to give blood samples. Some blood and chorion Factor VIII intron 22 Two probes (designated a 971 J Clin Pathol: first published as 10.1136/jcp.40.9.971 on 1 September 1987. Downloaded from

972 Tuddenham, Goldman, McGraw, Kernoff kb: 0 10 20 30 40 50 60 7 80 90 100 110 120 130 140 150 160 170 180 190 200 I I- I- I I II I I I I I I I I I I I I

I 23 4 56 789 101112 13 14 15AWAP 232425 26 I .. . ..a I 1 I a N I-1 5' 1 l -mNL -31i 1 3'

f t ? BcII Xbal Bgll Restriction fragment length polymorphismsin factor VIII

Fig Approximate location ofDNA polymorphisms detectable by restriction enzymes as indicated infactor VIII gene. Solid bars; exons numbered I to 2616, open bars; introns. Precise position of3' polymorphic BglI site has not been mapped.

and b) detect an XbaI polymorphism in intron 22 long enough to allow sufficient meiotic crossovers to (fig 1)." randomise the distribution of alleles on all chromo- Factor VIII 3' A fragment from the 3' end of factor somes. VIII cDNA that detects a BglI polymorphic site. The Table 2 gives haplotypes for BclI and XbaI. The precise location has not been mapped (fig 1).12 haplotype -/ + (large fragment/small fragment) was rare or absent. This means that women homozygous DNA MANIPULATION for BclI - are not informative for XbaI either Probes were gel purified or separated on CsC12 because they will be homozygous for XbaI -. Fortu- gradient by standard methods13 and labelled with nately, BclI + is the common allele, and such homo- [d-32P]dCTP by calf thymus DNA priming.'4 South- zygous women (60%) have a 37-5% (2 x 0 75 x 0-25 ern blotting"5 was performed after digestion of DNA x 100) chance of being heterozygous for XbaI. Table from a patient with appropriate restriction enzymes. 3 gives the haplotypes for BclI and BglI. The chance

of being heterozygous for both is reduced as the com- copyright. Results mon allele BclI + is associated with the common allele BglI +. Only those rare females who are homo- The reported incidence of the alleles for intragenic zygous for BclI + (4%) can get additional informa- and linked RFLP's varies somewhat according to tion from BglI. population and sample size. Table 1 gives percentages The net effect of these imbalances is that about based on our own series ofmainly northern European 70% of European females are informative-that is, ancestry. The largest discrepancy was the frequency of alleles at the factor VIII 3'BglI site'2 between http://jcp.bmj.com/ European and AfroCaribbean patients. In Europeans Table 2 Factor VIII BcJI-Xbal haplotypes about 80% of X chromosomes have this site-that is, 5 kb allele detected-in blacks the frequency is 74%, Haplotype No of Per cent of giving a higher expected heterozygosity rate for Bcll Xbal chromosomes chromosomes females. The cumulative heterozygosity for the three 0-8 4-8 (or 1 4) 45 61 intragenic RFLPs was not additive due to linkage dis- 0-8 6-2 15 20 1.1 4-8 (or 1-4) 0 0 equilibrium: the time elapsed from the mutation on September 29, 2021 by guest. Protected 1 1 6-2 14 19 events responsible for the DNA variations was not

Table 1 Expected Probe Enzyme* Alleles kb -/+ t Frequency (%) heterozygosity (%) Factor VIII exon 17-18 Bcl I 1 1/08 22/78 34 Factor VIII intron 22 XbaI 6 2/4 8 41/59 48 Factor VIII 3' Bgl I 20/5 20/80 32 DX13 BglII 58/2 8 50/50 50 Stl4 TaqI At least 8 > 90 *RFLP's are often referred to according to enzyme, such as factor VIII BclI. t - and + refer to absence or presence of restriction site yielding large or small alleles. J Clin Pathol: first published as 10.1136/jcp.40.9.971 on 1 September 1987. Downloaded from

Carrier detection and perpetual diagnosis ofhaemophilia A 973 heterozygous-at one or more polymorphic intra- polymorphic that more than 90% of females are genic sites. These women can be given firm genetic informative. It has the additional benefit of helping to advice based on linkage as the risk of crossover confirm or exclude paternity. between these markers and a putative mutation must DX13 was the first linked marker to be discovered average about 0-1% per meiosis. The total length of and is widely used. It is not very informative com- the factor VIII gene is about 200 kb or 02cM.16 The pared with Stl4 and is also subject to crossover maximum distance from marker to mutation (fig 1) is error.20 As several crossovers have affected both Stl4 130kb or about 0 - cM, equivalent to a 1/1000 cross- and DX13 they are most likely on the same side of the over rate per meiosis. factor VIII locus, which is unfortunate as no addi- For 30% of women who are non-informative with tional confidence is gained by combining them. intragenic RFLPs we must resort to linked markers. Stl4 and DX13 are at a small but important map Case histories distance from factor VIII. Table 4 gives crossover rates calculated from data sent to Dr I Peake as part In the family shown in fig 2 two sisters (III2 and I113) of an ongoing collection of data for the International have a 50% chance of carriership on pedigree. Based Society on Thrombosis and Haemostasis."7 These on the factor VIII Bcll alleles, II2 is informative with rates may be overestimates but most centres now the 0-8 allele being associated with haemophilia A in recheck antenatal diagnoses based on linked markers her father (I,) and son (III3). Therefore III2 cannot be with fetoscopic blood sampling. At least three ante- a carrier, whereas I113 is a carrier. This result is natal diagnoses based on Stl4 linkage have proved confirmed by the DX13 alleles. If III3 requires ante- erroneous due to crossover.18 19 The risk of error natal diagnosis in future, the factor VIII BclI poly- increases cumulatively with genetic distance from morphism could not be used as she is homozygous for "consultand" to affected index case. Despite these the 0-8 kb allele. DX13.BglII could be used as she is drawbacks Stl4 is a valuable marker as it is so Table 3 Factor VIII BcJI-BgJI haplotypes in northern 1 2 European population F Vill BclI 0.81 copyright. DX 13 2.81 No of Per cent of Bcl I Bgl I chromosomes chromosomes 1 2 08 5 29 66 0.81 0.811.1 08 20 0 0 2.8 1-1 5 5 11 5-81 15.8 1 1 20 10 23 1---- 322C)

Table 4 Factor VIII linkage (May 1986) 0-81 0.8111.1 0-81 0.8 http://jcp.bmj.com/ 2-81 558185.8 5.8 2.8 Meiosis Crossovers Per cent Fig 2 Family segregating haemophilia A. H112 and III3 DX 13/VIII intron 18 154 7 4 5 request carrier determination. F normal male; Ofemale; Stl4/VIII intron 18 133 6 4 5 * haemophilic male; 0D obligate carrier. 01, 1-1, RFL alleles in kb. on September 29, 2021 by guest. Protected

II

Fig 3 El normal male; Ofemale; E haemophilic male; (D obligate carrier. VIII:Cfactor VIII clotting activity; VWF:Ag Von Willebrandfactor antigen assayed in plasmafrom "consultand"1II3. J Clin Pathol: first published as 10.1136/jcp.40.9.971 on 1 September 1987. Downloaded from

974 Tuddenham, Goldman, McGraw, Kernoff

I.~~I Viltl,...... "vRl I 1.1 ni. .c i , . . v, 0 i.1.1 n.st F Vill Xba 1 6-2 4.A8 DX 13 Bgl If 2-8 5.8 2.8

0.8 0-8 1.1 1.1 6-2 6-211I 4.8 6-2 6-2 IfI 5 8 ff 2-8 12 8 2 8 5-3

6 2 2.8 Fig 4 O normal male; Ofemale; U haemophilic male; 0 obligate carrier. Kindred with sporadic case ofhaemophilia A. heterozygous, haemophilia A being associated with therefore be counselled on that basis and can have the 2-8 kb allele in this kindred. In that case the possi- first trimester fetal diagnosis using either the factor bility of crossover must be considered, giving about VIII BclI or the factor VIII XbaI RFLPs. This case 4% error risk for each fetus tested. shows the importance of maintaining the phenotype The family shown in fig 3 has three affected males assays (VIII:c and VWF:Ag) as combining informa- (14, T1, and II6) and two obligate carriers (12 and 13). tion of phenotype and linkage enhances diagnostic 113 requires carrier determination. Unfortunately, her potential. mother (12) is homozygous for all three RFLPs tested. The families in which a single case of haemophilia The phenotype tests for II3, however, show a very has occurred at time of consultation (so called spo-copyright. high probability (200: 1) that she is a carrier. She can radic cases) present a different problem-that of determining the level of the mutation, which can be 1 2 presumed to be of recent origin. Fig 4 illustrates the power of linkage analysis in such a kindred. A single haemophiliac III2 has the haplotype 11/6 2/2 8. This is identical with the haplotype of one X chromosome of his aunt 115 who was the "consultand." This X 1 ~~~2 3 chromosome, however, must have come from the hae- http://jcp.bmj.com/ mophiliac's normal grandfather I1. Therefore a new F Vil Xba I 4- mutation had occurred either at spermatogenesis 6.2114.8 from I, or at oogenesis from II2. In both cases IIs is excluded from carriership. Further information about 11

Carrier detection and perpetual diagnosis of haemophilia A 975 the 4-8 kb allele. As II, is heterozygous for this RFLP blood sampling at 18 weeks and terminated. The she can be offered antenatal diagnosis. patient was non-informative for factor VIII BclI but The family shown in fig 7 combine the problems informative for Stl4TaqI. The allele number 4 segre- addressed in the two preceding cases with a require- gates with haemophilia in this family and therefore ment for antenatal diagnosis of a male fetus (O). can be used for antenatal diagnosis, provided a pre- Comparison of alleles of I2 and 112 showed that 112 dicted normal male at 18 weeks' gestation is checked had inherited the 1 1 kb allele from her haemophilic by blood sample in case of error due to crossover. father (the phenotype results merely confirm that she The kindred analysed in fig 9 shows the benefit of is an obligate carrier of haemophilia A). A chorion using two independent intragenic markers. The con- biopsy specimen obtained at eight weeks' gestation sultand (II13) was pregnant when she requested diag- from III1 was shown to be male on karyotype. Fetal nosis. A haemophilic uncle (II7) was available for DNA was extracted5 and shown to bear the 0-8 allele. The pregnancy was therefore allowed to continue as the fetus would be normal. I In fig 8 the analysis was performed to help an obli- gate carrier (II12) in future pregnancies. A hae- mophilic fetus (IV1) was diagnosed by fetoscopic 11 F Vil Bcl I St 14 I F Vill Bcl I Taq Vill: C 37 u/dl III II VWF: Ag 100 u/dI

III copyright. IV

Fig 7 Malefetus . Antenatal diagnosis by DNA analysis Fig 8 Use ofSt14 TaqI RFLP. Intragenic marker despite non-availability ofaffected relative. non-informative. Severe haemophilia A two and one linked Linkage analysis using intragenic probe http://jcp.bmj.com/ 1 ~ 2,....2

1. 5 I

0.8 F Vil Bcl I (IVS 18) on September 29, 2021 by guest. Protected 6.2 F VIII Xba I (IVS 22) 5.8 DX 13 Bgl 11 I 4.- III

IV

Fig 9 ? Early conceptus, sex not determined. Illustrates exclusion ofcarriership requiring information from two intragenic RFLP's. J Clin Pathol: first published as 10.1136/jcp.40.9.971 on 1 September 1987. Downloaded from

976 Tuddenham, Goldman, McGraw, Kernoff testing but both grandparents (I1 and 12) were run the massive savings of preventing haemophilia deceased. The phase assignment for III3 was based on (table 6). The projected lifetime savings from efficient her father's haplotype. It was clear that the hap- prevention in one year would exceed the annual lotypes for 112 were such that she could not have United Kingdom outlay on all living patients. Can we inherited the haemophilic X chromosome from I,. afford not to offer this service as widely as possible? Therefore III3 is not a carrier and the pregnancy was allowed to continue undisturbed by any sampling. Gene probes were provided by Dr Richard Lawn of Genentech Inc, or supplied by the authors of the Discussion papers listed in the references. Availability of intragenic and linked RFLPs for gen- etic diagnosis of haemophilia A has greatly enhanced our ability to establish definitively carrier status. It is References now also possible to perform antenatal diagnosis by I Akhmetali MA, Aldorf LM, Alexanents S, et al. Methods for the linkage analysis using DNA extracted from chorion detection of haemophilia carriers: a memorandum. Bull WHO villus tissue as early as eight weeks' gestation. Several 1977;55:675. problems remain. As yet available intragenic markers 2 Mibashan RS, Rodeck CH, Thumpston JK, et al. Plasma assay are informative in only 70% of women. There is a of fetal factor VIIIc and IX for prenatal diagnosis of haemo- philia. Lancet 1979;i:1309- 1. need to identify more RFLPs within and flanking the 3 Wright AF. DNA analysis in human disease. J Clin Pathol factor VIII gene. To date, only about 15% ofthe gene 1986;39:1281-95. has been screened for polymorphism,8 11 and it is 4 Bell GI, Kavam JH, Rutter WJ. Polymorphic DNA regions highly likely more remain to be discovered. As fami- adjacent to the 5' end of the human insulin gene. Proc Natl Acad Sci USA 1981;78:5759-63. lies are often incomplete and some haemophiliacs die 5 Williamson R, Eskdale J, Coleman DV, Niazi M, Loeffler FE, young it is essential to retain the phenotype assays, as Modell BM. Direct gene analysis of chorionic villi: a possible pointed out by Graham et al.21 technique for first-trimester antenatal diagnosis of haemo- Certain difficulties of linkage analysis are inherent globinopathies. Lancet 1981;ii:1125-7. 6 Davies KE, Young BD, Elles RG, Hill ME, Williamson R.copyright. and can only be contained by vigilance and attention Cloning of a representative genomic library of the human X to detail, bearing in mind that paternity may not chromosome after sorting by flow cytometry. Nature always be as stated. Table 5 lists some of these prob- 1981;293:374-6. lems. In addition, the costs ofanalyses entailing DNA 7 Harper K, Winter RM, Pembrey ME, Hartley D, Davies KE, Tuddenham EGD. A clinically useful DNA probe closely manipulation are high. Against these costs should be linked to haemophilia A. Lancet 1984;ii:6-8. set the considerable improvement in the genetic coun- 8 Gitschier J, Drayna D, Tuddenham EGD, White RL, Lawn RM. selling service to potential carriers, and in the long Genetic mapping and diagnosis of haemophilia A achieved through a Bcl I polymorphism in the factor VIII gene. Nature 1985;314:738-40. http://jcp.bmj.com/ 9 Oberle I, Drayna D, Camerino G, White R, Mandel J-L. The Table 5 Some problems oflinkage analysis telomeric region of the human X chromosome long arm: Pres- ence of a highly polymorphic DNA marker and analysis of I Clerical error, sample misidentification etc recombination frequency. Proc Natl Acad Sci USA 2 Non-paternity 1985;82:2824-8. 3 Non-availability of family members 4 Unknown level of new mutation 10 Oberle I, Camerino G, Heilig R, et al. Genetic screening for 5 Technical faults ofpartial restriction, pipetting error, probe Haemophilia A (Classic haemophilia) with a polymorphic contamination DNA probe. N Engl J Med 1985;312:682-6. 6 Length ofanalysis in urgent cases 11 Wion KL, Tuddenham EGD, Lawn RM. A new polymorphism on September 29, 2021 by guest. Protected 7 Non-informative RFLP in the factor VIII gene for prenatal diagnosis ofhaemophilia A. 8 Crossover Nucleic Acids Res 1986;14:4535-42. 12 Antonarakis SE, Waber PG, Kittur SD, et al. Haemophilia A. Detection of molecular defects and of carriers by DNA Table 6 Economics ofsevere haemophilia care and analysis. N Engl J Med 1985;313:842-8. prevention per patient per year (£) 13 Maniatis T, Fritsch EF, Sambrook J. Molecular cloning. A labo- ratory manual. New York: Cold Spring Harbor Laboratory, Care: 1982. Medical and social services 600 14 Taylor JM, Illmensee R, Summer J. Efficient transcription of Factor concentrate, 35 000 units at 15p per unit 5250 RNA into DNA by avian sarcoma virus polymerase. Biochim Lifetime cost assuming life expectancy of65 years 373 750 Biophys Acta 1976;442:324-30. United Kingdom annual cost 11 500000 15 Southern EM. Detection of specific sequences among DNA electrophoresis. J Mol Biol Prevention: fragments separated by gel No ofpatients born with severe haemophilia in United 1975;98:503-18. Kingdom 50 16 Gitschier J, Wood WI, Goralka T, et al. Characterization of the Preventable (not sporadic) cases 33 human factor VIII gene. Nature 1984;312:326-30. Lifetime cost = potential saving £12 300 000 17 Peake IR, Bloom AL. Recombination between genes and closely linked polymorphisms. Lancet 1986;i:1335. J Clin Pathol: first published as 10.1136/jcp.40.9.971 on 1 September 1987. Downloaded from

Carrier detection and perpetual diagnosis ofhaemophilia A 977 18 Driscoll MC, Miller CH, Goldberg JD. Aledort LM, Hoyer LW, 21 Graham JB, Green PP, McGraw RA, Davis LM. Application of Golbus MS. Recombination between Factor VIII:c gene and molecular genetics to prenatal diagnosis and carrier detection St14 locus. Lancet 1986;ii:279. in the haemophilias: some limitations. Blood 1985;66:759-64. 19 Lehesjoki A-E, de la Chapelle A, Rasi V. Haemophilia A: two recombinations detected with probe Stl4. Lancet 1986;ii:280. 20 Winter RM, Harper K, Goldman E, et al. First trimester prenatal Requests for reprints to: Dr EGD Tuddenham, Haemostasis diagnosis and detection of carriers of haemophilia A using the Research Group, Clinical Research Centre, Watford Road, linked DNA probe DX13. Br Med J 1985;291:765-9. Harrow, Middlesex HAI 3HJ, England. copyright. http://jcp.bmj.com/ on September 29, 2021 by guest. Protected