J . Cell Sci. Suppl. 4, 445-458 (1986)Printed in Great Britain © The Company of Biologists Limited 1986

445

THE M OLECULAR GENETICS OF HAEMOPHILIA

A AND B

G. G. BROWNLEE

S ir William Dunn S chool o f Pathology, U niversity o f Oxford, South Parks Road, Oxford 0X1 3RE, UK

I NTRODUCTI ON

Haemophilia has been known since biblical times as an inherited bleeding condition, as boys born into families known to have the disease were excluded from ritual circumcision in the third century a d (quoted by McKee, 1983). The pattern of inheritance whereby males were affected, whereas females were not (although they could transmit the disease to future generations), intrigued the 19th century biologists and was not adequately explained until the genetic basis of sex was understood. The clinical symptoms of the disease can be very serious and were life- threatening before replacement therapy was available. Internal bleeding occurred into muscles and joints, often without obviously following any trauma and patients could die of massive internal haemorrhage. Haemophilia A and B pose identical clinical pictures and both show a similar X-linked pattern of inheritance. Their differential diagnosis depends on laboratory clotting tests and they were first clearly distinguished in 1952. Subsequent work showed that the former condition involved defects in the protein factor VIIIC and the latter in factor IX. Both proteins function in the middle phase of the intrinsic clotting cascade (Fig. 1). Haemophilia A is the commoner disease, occurring in Caucasians in approximately one in 5000 males, and is also referred to as classical haemophilia. Haemophilia B occurs in approximately 1 in 30000 males and is also known as Christmas disease, after the name of a patient, Stephen Christmas, in whom the disease was characterized early on.

Between 1960 and 1980 the methods of analysing the sequence of amino acids in proteins that had been pioneered by Sanger and his colleagues on insulin were applied to the proteins of the clotting cascade, including the factor IX protein. By1979 the entire amino acid sequence of the bovine factor IX was known (Katayama et al. 1979) and there was evidence suggesting that a rather similar sequence might be present in humans. This was a considerable feat of work because of the low yield (about 5 jUgml-1 of the protein) in plasma. But this basic knowledge of protein structure was not as easily obtained for factor VIIIC, partly because of its minute yield in plasma, but also because its high molecular weight and instability made it particularly difficult to characterize. In fact, factor VIIIC exists as a minor component of the high molecular weight complex referred to as factor VIII, and the bulk of this complex is composed of another protein, the von Willebrand factor.

446 G. G. Brownlee

In this short review, I will briefly describe and discuss the advances made since1980 that have enabled the genes for factor VIIIC and IX to be isolated by recombinant DNA methods. I will then describe progress in the uses of such gene probes in clinical diagnosis of carriers, as well as their use for studying the molecular pathology of the disease and the production of genetically engineered protein.

CLONING AN D CHARACTERIZATION OF THE GENES FOR FACTORS VIIIC AND IX

In the period between 1970 and 1980, considerable advances were made in the purification of messenger RNA from highly specialized cells, such as the a and ¡3 globin mRNA from reticulocytes. It became possible to clone these highly enriched mRNA preparations, first by making complementary DNA, or cDNA, copies of the mRNA, and then subsequently converting this material to double-stranded DNA in vitro and cloning in plasmid or in phage A vectors propagated in the bacterium, Escherichia coli. Subsequently, these cDNA clones could be characterized and used to isolate and study the gene organization. For clotting factor IX, a different approach was necessary as its messenger RNA was present in extremely low concentrations in liver. Nevertheless, using the approach of synthesizing short oligonucleotide probes based on a knowledge of the amino acid sequence of portions of the protein, the problem became soluble. We synthesized oligonucleotide probe mixtures for several regions of the bovine factor IX sequence to enable the isolation

Surface contactIXII Xlla

XI- -Xla

Ca

IX-

X-

IXaCa • VIII tissue extractPhospholipid y j[ c a+ +

-»Xa*- -X

II-

Ca1 Phospholipid

■HaI -Fibrin

Fig. I. The clotting cascade. The intrinsic pathway is shown from top left to bottom right. The extrinsic pathway (activation of X by activated VII) is also shown. This is a simplified version showing the main features, but it omits feedback loops and a step, e.g. factor V ila activation of IX, interconnecting the two pathways (from Austen & Rhymes, 1975, with permission).

Haemophilia A and B 447

of an initial short cDNA probe from bovine liver (Choo, Gould, Rees & Brownlee, 1982). Others used essentially the same principles, although the details differed slightly (Kurachi & Davie, 1982; Jaye et al. 1983). These initial cDNA clones were then used as probes in order to isolate both human genomic and cDNA clones establishing the nature of the gene structure as about a 34x 103 base long region of DNA split into eight coding regions or exons of varying length, which are inter­rupted by seven non-coding regions or introns (Anson et al. 1984). Subsequently, the entire gene sequence has been established by Davie and co-workers (Yoshitake et al. 1985).

For factor VIIIC a variation on the method of using oligonucleotide probes was used. Because there was some uncertainty as to which tissue to use as a source of mRNA in the preparation of factor VUIC-containing cDNA libraries of clones, both groups who successfully cloned and characterized the factor VIIIC gene chose to isolate the initial clone from a short section of the factor VIIIC gene. It turned out to be extremely long and complicated and is about 186X103 bases in length, and boasts 26 introns. The mRNA codes for a mature glycoprotein of 2332 amino acids in length. This gene is the largest known to date and occupies approaching 0-1 % of the total length of the X chromosome (Gitschier et al. 1984; Toole et al. 1984).

D I A G N O S I S OF C A R RI E R S AN D P R E N A T A L D I A G N O S I S OF AF FECTED

I N D I V I D U A L S

When girls in known haemophiliac families reach child-bearing age, they naturally wish to know the risk of having an affected son. Girls have, on average, a 50 % chance of carrying the defective gene, assuming they inherit this from their mother. Traditional methods of carrier detection in such girls rely on the measurement of the concentration of factors VIIIC and IX in plasma. On average, the value in carriers is half that in normal individuals, but unfortunately the diagnostic value of such measurements is limited because of the wide spread of actual values among both normal individuals and carriers. The range of values in carriers is explained by lyonization, i.e. the random inactivation of one X chromosome early in embryonic life, resulting in the fact that the factor-VIII- or IX-producing cells in an individual carrier may be strongly biased in favour of one or other of the paternal or maternal chromosomes.

Although of some value, these classical methods of diagnosis of carriers are thought to have an error-rate of, on average, 17 % in haemophilia A and a somewhat similar rate in haemophilia B (see review by Giannelli, 1986; Barrow, Miller, Reisner & Graham, 1982; Graham, Flyer, Elston & Kasper, 1979). Prenatal diagnosis of affected males, in contrast, is reliable so long as a foetal blood sample can be obtained from the foetal circulation (umbilical vein) uncontaminated with maternal blood (see Mibashan, Giannelli, Pembrey & Rodeck, 1986). A simple clotting assay clearly distinguishes affected from normal individuals. The disadvantage of this procedure is that it requires a skilled gynaecologist to carry out this operation. Furthermore, foetal blood sampling cannot easily be carried out before the 19th or 20th week in

448 G. G. Brownlee

pregnancy, when a termination, if desired, is a reasonably unpleasant experience for the mother.

Cloned DNA probes derived from parts of the gene for factors VIIIC or IX offer an alternative and more reliable method of diagnosing carriers. Such clones can also be used for prenatal diagnosis at an early stage in pregnancy, using chorionic trophoblast or amniotic cell samples. Such procedures have been pioneered for the antenatal diagnosis of sickle cell anaemia and some forms of thalassaemia. In these cases, gene probes can identify the mutation directly causing the gene defect (reviewed by Weatherall, 1982). In contrast, haemophilia A and B gene defects are not fixed and maintained in the population by virtue of heterozygote advantage (as is believed to be the case for sickle cell anaemia). Hence mutations are believed (see below) to be caused by a variety of defects of recent and independent origin. For this reason, it is not practical to develop mutation-specific probes for every affected pedigree, so that an indirect diagnosis relying on genetic linkage to a polymor­phism identified within or very close to a gene becomes the method of choice. Polymorphisms in DNA can be detected by different methods, but those that result in the appearance of one or other of two bands of differing size in a restriction enzyme digest of DNA on an agarose gel are the easiest to work with experimentally. These polymorphisms are referred to as restriction fragment length polymorphisms (RFLP). Three distinct such RFLPs have been identified and used for carrier diagnosis in haemophilia B pedigrees (Giannelli et al. 1984; Winship, Anson, Rizza & Brownlee, 1984; Giannelli, 1986). Two such polymorphisms are known at present for haemophilia A (Gitschier et al. 1985a; Antonarakis et al. 1985).

One current limitation of the linkage analysis with RFLPs is that the frequencies of these polymorphisms in the general population, which appear to mimic that in the affected pedigrees, is such that carrier diagnosis is possible as yet on an average of only 68 % of known Caucasian haemophilia B pedigrees and about 50 % of Caucasian haemophilia A pedigrees. Moreover, as the assays are indirect, they rely on the availability of blood samples not only from the potential carrier under investigation but also from a known patient within the pedigree, preferably a brother, as well as from the mother and father. Fig. 2 illustrates the example of such an RFLP linkage study taken from Winship et al. (1984). The person under investigation III3 was shown to have a genotype aiai. Given the genotypes of her mother (a ^ ) , father (ai) and affected brother (ai), we concluded she must be a carrier. If she had been aia2, we would have excluded her as a carrier. The chances of error in this method, omitting issues like paternity and technical errors in the assay due to mixing up samples, are the chance that recombination occurred in the female oocyte at mei- osis between the actual disease mutation (somewhere in the factor IX gene) and the RFLP (also located and at a known position within the factor IX gene). Recombination probably occurs non-randomly in chromosomes and we do not know whether this is more frequent or less frequent than average near the factor IX locus; but for the purposes of argument, if we assume there is a possibility of three cross­overs per X chromosome and that the X chromosome is 200X106 bases long, and

Haemophilia A and B 449

1Ial

o-

1

-©

2o

©

□

4□ 5□ I

I

E e2 m2 m3 I,_______11________ ii________ i i________ ;i----------

mm tBtB(11.5 )

sr • •Fig. 2. Carrier diagnosis (below) of a haemophilia B pedigree (above) using an XmnI RFLP in the factor IX gene and a subgenomic probe VIII that gives ai ( 1 1 - 5 x 1 03 base) and a2 (6-5X 103 base) alleles (from Winship et al. 1984, with permission).

knowing the maximum distance between the polymorphism and gene defect cannot be more than 30X103 bases, there will be an even chance of the linkage being broken in the ratio of 3 0 x 3 : 2 0 0 x l 0 3 or —1 in 2 x l 0 3 times. In percentage terms this is 0-05% chance. For such reasons, we concluded that the indirect linkage analysis w ith intragenic RFLPs is 99-9% accurate for factor IX (Winship et al. 1984). For factor VIIIC with a gene of approximately five times the length of the factor IX gene, the corresponding value is 99-7% . In the clinical situation, where families and

450 G. G. Brownlee

carriers may be counselled, one must remember that this is an average probability and only an approximate calculation, and that there must be a chance, as more families are counselled as a result of these new methods of analysis, that cases will arise in which the linkage is broken and therefore a wrong diagnosis is made.

More distantly related RFLP markers, e.g. anonymous DNA probes, have been advocated by some experimentalists for carrier and prenatal diagnosis, particularly in the case of factor VIII (Gitschier et al. 1985a) in which few intragenic RFLPs are known to date. However, a problem with these markers, e.g. DX13 and S tl4 , at a close, but still uncertain (Janco et al. 1986), distance from the disease locus, is that the recombination fraction is not known with sufficient accuracy to make them as reliable diagnostically as specific gene probes. Only if and when 200 to 300 total meioses are studied and the recombination fraction is shown to be less than 1 % would I advocate their use clinically.

If obligatory carriers fail to show heterozygosity for any of the known gene-specific polymorphisms, of course, no accurate information is forthcoming for potential carriers - a situation hardly reassuring for the family under investigation, as they would have to fall back on classical diagnostic methods or somewhat less-reliable diagnostic markers such as DX13 or S tl4 in the case of haemophilia A. One may calculate that if 95 % of all families with haemophilia A or B are to be counselled, a knowledge of at least five different gene-specific RFLPs will be required even if we assume that (1) each RFLP is distributed in the haemophiliac population at the theoretical maximum frequency of 50% , and (2) each RFLP segregates indepen­dently. In practice, not all RFLPs are favourably distributed and because of the tendency of adjacent regions of DNA to be inherited in a linked way some of the RFLPs will not be distributed independently of each other in the population at large. This linkage disequilibrium can severely reduce the value of additional RFLPs (as shown for the XmnI polymorphism in factor IX; Winship et al. 1984) when information from a Taql RFLP is already known. Because of such difficulties, which have also been described for the /3-globin locus (Weatherall, 1982), a realistic estimate of the number of RFLPs required in order to ‘catch’ 95 % of the affected pedigrees is nearer seven to ten. This value will be difficult to achieve in the short term, but in the medium term, say the next 3 -5 years, it should be possible. Of course, a highly polymorphic ‘minisatellite’ type of sequence such as that occurring near the ar-globin locus (Weatherall, Higgs, Wood & Clegg, 1984) would be ex­tremely valuable, but to date there is no evidence of this near the factor VIIIC or IX loci. The current position on factor IX is that a fourth RFLP involving an M nll RFLP is under investigation in Oxford (P. R. Winship, personal communication). Similarly, I am sure that further RFLPs will be found in the much longer factor VIIIC gene. So far, only a few prenatal diagnoses have been performed on foetal chorionic villi or amniotic samples using RFLPs in haemophilia A and B, but as the modern techniques for carrier diagnosis become more available and potential carriers in affected families find out their status, I predict an increasing demand for such antenatal information.

Haemophilia A and B 451

M O L E C U L A R DEF ECTS IN H A E M OP H I LI A A A N D B

Evidence of heterogeneity exists in both haemophilia A and B as there is variation in the clinical severity of the disease in different pedigrees, as well as variation in laboratory tests of clotting activity and the antigen concentration measured in samples of blood taken from different patients. We therefore expect a wide range of molecular defects and indeed a precedent for this exists in the variety of mutations already known in the haemoglobin disorders (Weatherall et al. 1984). Haemophilia patients may be conveniently subdivided into those patients in whom protein is present as detected by immunological methods, referred to as antigen positive, and those in whom it is absent, referred to as antigen negative. The former might be expected to have point mutations in that region of the gene encoding amino acid residues of the protein and studies of such patients can clearly pinpoint critical functional regions. The latter are more likely to be mutants involving deletions of reasonably large regions of the molecule or critical point mutations involved in the biosynthesis of mRNA (e.g. splice junction sequences), or of protein. Critical point mutations involved in protein synthesis might be changes in the translation initiation or termination codons, or indeed the introduction of such nonsense codons in incorrect coding positions, or mutations affecting protein secretion and maturation. A small subgroup of patients of the antigen negative subgroup are referred to as ‘inhibitor’ patients because they have specific anti-factor-VIIIC or -IX antibodies in their plasma, which arise in response to therapy with injected normal clotting factors.

Only a few defects of patients have been characterized fully at the molecular level using recombinant DNA methods and the results are summarized in Table 1. (For completion, I include one factor IX antigen positive patient, factor IX ChapeiHiii (Noyes et al. 1983), whose defect has been characterized by a study of the abnormal protein using the methods of amino acid sequencing.) An antigen positive patient, haemophilia Boxford3 i has been investigated by recombinant DNA methods (Bentley et al. unpublished data) and this patient possesses a point mutation at.an amino acid in the propeptide precursor domain of the protein, specifically at amino acid —4, (i.e. 4 amino acids preceding the N-terminal tyrosine of the mature factor IX ), which results in an abnormally long factor IX protein with an N-terminal extension of 18 amino acids accumulating in plasma. This material is inactive in clotting for reasons that are unknown. Nevertheless, it clearly shows the need for accurate protein processing, which is obviously critically dependent on the arginine residue at amino acid —4 and defines a processing intermediate that has not previously been charac­terized. Another antigen positive patient, factor IXAiabama> has been characterized as a change at amino acid 47, which affects calcium binding (Davis et al. 1984, and personal communication). Among the factor IX antigen negative patients, two patients, haemophilia B 0 x f 0 r d i a n d 2 , from independent pedigrees have been shown to have a defect in different splice—donor junction sequences. One is at the G-T donor at the 3' end of exon f (exon f is the sixth of the 8 coding regions of the gene), which is changed to T-T (Fig. 3; and Rees, Rizza & Brownlee, 1985). The other involves the G-T at the 3' end of exon c (the third of the 8 coding regions of the gene), which

Table 1. Characterized molecular defects in haemophiliaHaemophilia B Haemophilia A

Subgroup Patient Defect Patient DefectAntigen positive Factor I X A l a b a m a

Factor I X C h a p e lH i l l

Haemophilia B0xfc,rd 3

Asp—> Gly (47) Arg—» His (145) Arg—» Gin ( -4 )

Antigen negative Haemophilia B 0 x f 0r d 1 G-T —» T-T, donor splice junction of exon f

H22 Arg (2307) —» TGA, chain termination giving truncated protein

Haemophilia B 0 x f 0r d 2 G-T —> G-G, donor splice junction of exon c

H51 Deletion of 22X103 bases of last exon (no. 26) and extending 3' to the gene, giving truncated (2282 amino acid long) protein

Inhibitors 4 patients from independent pedigrees in UK

Partial and complete gene deletions; exact length of deletion not known

Family A

Family B

H2

H96

80 X103 base deletion within gene

Arg (1941)—» TGA, chain termination giving truncated protein

Arg (2209) — » TGA, chain termination giving truncated protein

Deletion of 39X103 bases within gene, excising exons 23—25

The numbers in parentheses refer to the amino acid residues of the protein.

452 G. G. Brow

nlee

Haemophilia A and B 453

is changed to G-G (P. R. Winship, unpublished). These mutations are very reminiscent of the kind of defect that occurs in the /3-globin gene in /3° thalassaemias (Weatherall et al. 1984) and it is instructive to note how critical a single point mutation in a vital processing pathway can be, causing as it does clinically severe haemophilia in both affected patients.

The third group of patients, which have been studied in detail in haemophilia B and less extensively in haemophilia A, are those that produce inhibitors. We argued that the reason these patients made antibodies was that their immune system had not been made tolerant to normal factor IX and therefore such patients were likely to have gene deletions. Of course, critical point mutations involving RNA processing or nonsense mutations affecting translation may also have the same effect as gene deletion by preventing normal protein synthesis. Of the six known UK patients, five had evidence of partial or complete gene deletion (Giannelli et al. 1983; Peake, Furlong & Bloom, 1984). Presumably, the one non-deletion inhibitor patient described by Giannelli et al. (1983) has a critical point mutation elsewhere. Other examples of factor IX inhibitor patients are under study and three Italian patients seem to have at least partial gene deletions whereas an Australian patient may be of the non-deletion type. In haemophilia A, four inhibitor patients have been characterized in detail. Only one (family A, Table 1) has an extensive gene deletion estimated as greater than 80X103 bases. The others either have much shorter deletions near the 3' end of the gene (patient H96) or point mutations generating an aberrant stop codon, which would effectively produce a truncated factor VIIIC protein in family B and patient H2 (Gitschier et al. 19856; Antonarakis et al. 1985). The postulate still remains valid, however, that the basic reason for the development of the inhibitor status, a clinical complication second only in importance to AIDS contamination of concentrates, is the same in both haemophilia A and B, i.e. that the patient’s tolerance has not been induced to critical epitopes of the protein (Giannelli et al. 1983). However, the position of these critical epitopes in the protein, and the probable differences in detail in the development of tolerance to factor VIIIC and IX remain unknown. Two other clinically severe haemophilia A patients H22 and H51 (see Table 1) have been characterized at the molecular level. One is caused by another point mutation, generating an aberrant translation stop codon, and a second is caused by a short deletion at the 3' terminus of the gene (Gitschier et al. 19856). Neither is an inhibitor patient, but Giannelli & Brownlee (1986) note that the probable effect of these latter two mutations on the protein would be expected to be

----------------------------E x o n f ------------------------------------------------------- ►190

V V G G E D A K P G Q F P W QTTGTTGGTGGAGAAGATGCCAAACCAGGT CAATT CCCTTGGCAGGTACTTTATACTGAT GGTGTGTCAAA

2 11 30 2 1 1 4 0 21 15 0 21160 ^ 2 1 1 7 0 21 18 0 21 19 0

T

Fig. 3. Sequence surrounding the 3' end of exon f of the normal factor IX gene showing the G—»T mutation in haemophilia B0xf0rdi. One-letter symbols are used for the amino acids (from Rees et al. 1985, with permission).

454 G. G. Brownlee

Xq

2 5 - _ _

_ HPRT26

FIX27 [_J FS

28 * ------- ̂ HEMAG6PD

CBALD

Fig. 4. Diagram of the terminal section of the long arm of the human X chromosome showing some disease loci. For abbreviations, see the text. HPRT is the locus for hypoxanthine guanine phosphoribosyl transferase.

less than the mutations discussed above in the two patients with inhibitors. To date no patients have been described with defects in the promoter or in the A-A-U-A-A-A polyadenylation region of the gene.

Further studies of antigen positive patients should give valuable information about the critical regions for function in both the factor VIIIC and the factor IX molecules. Factor IX and its activated form IXa have to interact with at least five other molecules in the middle stage in the intrinsic clotting pathway, so we must expect many parts of the molecule to be critical for function and, or, correct folding of the protein. Sim ilar arguments apply to factor VIIIC , although in this case we sus­pect that the central carbohydrate-rich portion of this very large molecule is less important functionally than the rest of the molecule, as it has diverged extensively in amino acid sequence among mammals (Orr et al. 1985).

R E G I O N A L L O C A L I Z A T I O N ON THE X CH R OM OS O ME

Fig. 4 shows the localization of the factor IX locus (labelled FIX) to band q27 near to the tip of the long arm of the X chromosome. The localization was discovered using cloned factor IX probes (e.g. see Boyd et al. 1984) and has been more recently refined to band q27.1 by in situ hybridization to extended early metaphase chromosomes (V. J . Buckle, personal communication). The short cytogenetic distance from the locus at q27.3 for mental retardation with macro-orchidism associated with fragility (FS) is clear. Unfortunately, however, there is still too much recombination (20 %) between the factor IX locus and the FS for the factor IX probes to have any real value in carrier diagnosis of this rather common and

Haemophilia A and B 455

depressing disease (e.g. see Choo et al. 1984). The haemophilia A locus (HEMA) is known to be closely linked to a group of three other markers; glucose-6-phosphate dehydrogenase (G6PD), colour blindness (CB) and adrenoleukodystrophy (ALD), but again the recombinational fraction between haemophilia A and FS is too great for the factor VIII probes to be useful in diagnosis of the mental retardation syndrome. Many other anonymous gene markers are now known for this region of the X chromosome, but those that have been tested so far are no closer to the FS locus than factor IX. A closely linked probe or a specific gene probe is urgently needed.

G E N E T I C A L L Y ENGI NEERED FA CTO RS VIIIC A N D IX

Haemophilia A and B patients need regular injections of factors VIIIC and IX if haemorrhage is to be avoided and controlled. Unfortunately, there is risk of viral contamination introduced in the donor blood used for these preparations, which has not been completely removed in the purification procedure. Recently, AIDS has superseded hepatitis B and the non-A, non-B hepatitis as the most hazardous complication of therapy. Although the AIDS virus is fortunately rather heat-labile, we do not yet know how successful the heat treatment (instituted in 1985) will be and whether other viruses might appear from time to time that cannot be so easily inactivated. It is therefore highly desirable to produce the required proteins from a genetically engineered source non-contaminated with viruses. This should in addition give a well-standardized product and manufacturers would no longer need to rely on blood donors for their starting material.

The ability to produce factor VIII in vitro from cloned DNA was first described late in 1984 by two genetic engineering companies (Wood et al. 1984; Toole et al. 1984). They introduced modified cDNA clones into mammalian cells, i.e. kidney cells or T lymphoma cells, and showed that biologically active factor VIIIC was secreted into the medium, as assayed by highly specific in vitro tests. In 1985, three papers, including one from my own laboratory, reported similar successes with factor IX, also in mammalian cells (Anson, Austen & Brownlee, 1985; de la Salle et al. 1985; Busby et al. 1985). Both liver and kidney cells successfully produced active material, although the specific activity of the factor IX protein was higher in the liver cell, presumably reflecting the fact that this cell tissue had correctly modified and processed the factor IX. As liver is the tissue in which factor IX is normally synthesized, it might be expected to be appropriate for correct expression and pro­duction of factor IX that is indistinguishable from the raw material. Nevertheless, in all these studies, of both factor VIIIC and IX, minute yields of product are reported. One of the best yields of approaching 1 /¿g of factor IX per ml of medium was reported in one of the papers on factor IX (Busby et al. 1985). Even with this, I estimate it would be necessary to culture hundreds of thousands of litres of cells to obtain enough material for the U K requirement of 50 g of factor IX for one year. Given that purification of proteins on a large scale necessarily entails losses, it is clear that more efficient small-scale synthesis has to be developed before an investment in a large-scale industrial process is made. Fortunately, there is an indication that

456 G. G. Brownlee

a protein related to factor IX (the anti-coagulant protein C) can be successfully produced in high yields in Chinese hamster ovary cells using recombinant DNA methods. In these cells the protein C gene copy number can be amplified by linkage of the clone to the gene for dihydrofolate reductase followed by growth of cells in methotrexate. Thus it may be possible to produce cell lines giving at least 10 times more factor IX. For factor VIIIC, there is a similar requirement for maximization of the yields of genetically engineered material on a small scale.

I estimate that it will be 3 -5 years before the necessary industrial processes for factors VIIIC and IX will be developed and the material adequately characterized and tested for clinical application to human beings. Nevertheless, the product should be safer, free of viruses, and, I suggest, will supersede traditional products made from blood with all their inherent problems.

I thank the Medical Research Council for support.

R EFER EN CES

A n so n , D. S ., A usten , D. E. G. & Brownlee, G. G. (1985). Expression of active human clotting factor IX from recombinant D N A clones in mammalian cells. Nature, Lond. 315, 683-686.

A nson , D. S ., Choo, K . H ., Rees, D. J. G ., G iannelli, F ., G o u ld , K . , Huddleston , J. A . & Brow nlee , G. G. (1984). Gene structure of human anti-haemophilic factor IX. EMBO J . 3, 1053-1064.

A n ton arakis, S . E., W aber , P. G., K ittur, S . D ., Patel, A . S ., K a z a z ia n , H. H., J r , M ellis , M . A ., C ounts, R. B ., S tamatoyannopoulos, G., Bow ie , E. J . W ., F a s s , D . N ., Pittman, D . D ., W osney , J . M . & T oole, J . J . (1985) . Hemophilia A : detection of molecular defects and of carriers by D N A analysis. N. E ng.J . Med. Oct. 3, 8 4 2 -8 4 8 .

A usten , D. E. G. & R hymes, I. L. (1975). A L aboratory M anual o f B lood Coagulation. Oxford: Blackwell Scientific.

Ba r r o w , E. S ., M iller , C . H., Reisner, H. M . & G rah am , J. B. (1982). Genetic counselling in haemophilia by discriminant analysis 1975-1980. .7- m ed. Genet. 19, 26.

Boyd , Y . , Buc kle , V . J ., M unro , E. A., C hoo, K . H., M igeon, B. R . & C r aig , I. W . (1984). Assignment of the haemophilia B (factor IX) locus to the q26—qter region of the X chromosome. Ann. hum. Genet. 48, 145-152.

Bu sb y , S., K um a r , A., J oseph , M., Ha lfpap , L., In sley , M., Berkner, K . , K urach i, K . & W oodbury, R. (1985). Expression of active human factor IX in transfected cells. Nature, Lond. 316, 271-273.

C hoo, K . H ., G eorge, G ., F ilby , G ., Ha llid a y , J . L ., L eversha, M ., W ebb, G . & D a n k s ,D . M . (1984). Linkage analysis of X-linked mental retardation with and without fragile X using factor IX gene probe. L ancet i i Aug., 349.

C hoo, K. H., G o u ld , K. G ., Rees, D. J. G . & Brownlee , G . G . (1982). Molecular cloning of the gene for human antihaemophilic factor IX. N ature, Lond. 299, 178-180.

D a v is , L . M ., M cG r a w , R . A ., G rah am , J . B ., Roberts, H. R . & S tafford , D . W . (1984 ) . Identification of the genetic defect in factor IXAiabama' D N A sequence reveals a G ly substitution fo r A sp 46 . B lood 64 (Sup p l. 1), 262a.

de l a S a l l e , H ., A ltenburger, W ., Elkaim , R ., D ott, K ., D ieterle, A ., D rillien, R ., C a se n a v e , J.-P., T olstoshev, P. & L ecocq, J.-P. (1985). Active y-carboxylated human factor IX expressed using recombinant D N A techniques. Nature, Lond. 316, 268-270.

G ian nelli, F . (1986) . The contribution of gene specific probes to the genetic counselling of families segregating for haemophilia B (factor IX deficiency). In B ari Int. Conf. on F a cto r VIII/von W illebrand F a c to r -B io lo g ica l and C linical A dvances (ed. N. Ciavarella, Z. Ruggieri & T. S. Zimmerman). Milan: Wichtig (in press).

G ian nelli, F . & Brownlee , G . G . (1986). Nature, Lond. (in press).

Haemophilia A and B 457

G i a n n e l l i , F . , C h o o , K . H ., R e e s , D . J . G., B o y d , Y., R iz z a , C . R . & B r o w n le e , G. G. (1983). Gene deletions in patients with haemophilia B and anti-factor IX antibodies. Nature, Lond. 303, 181-182.

G ia n n e l l i, F . , C h o o , K . H ., W in sh ip , P . R ., A n so n , D . S., R e e s , D . J . G . , F e r r a r i, N., R i z z a , C. R . & B r o w n l e e , G . G . (1984). Characterisation and use of an intragenic polymorphic marker for detection of carriers of haemophilia B (factor IX deficiency). L ancet i, 239-241.

G it s c h i e r , J ., D r a y n a , D., T u d d e n h a m , E. G . D., W h ite , R. I. & L a w n , R. M. (1985a). Genetic mapping and diagnosis of haemophilia A achieved through a Bell polymorphism in the factor VIII gene. Nature, Lond. 314, 738-740.

G it s c h i e r , J ., W o o d , W . I., G o r a l k a , T. M., W io n , K. L ., C h e n , E . Y., E a to n , D. H., V e h a r , G . A . , C a p o n , D. J. & L a w n , R. M. (1984). C h arac te riza tio n o f th e h u m an fa c to r V I I I g en e. N ature, Lond. 312, 326-330.

G itsch ier , J ., W o o d , W . I., T u d d e n h a m , E. G . D., S h u m a n , M. A ., G o r a l k a , T . M., C h e n ,E. Y. & L a w n , R. M. (19856). Detection and sequence of mutations in the factor VIII gene of haemophiliacs. Nature, Lond. 315, 427-430.

G r a h a m , J. B., F l y e r , P., E ls t o n , R. C. & K a s p e r , C. K . (1979). Statistical study of genotype assignment (carrier detection) in haemophilia B. Thromb. Res. 15, 69.

J a n c o , R . L . , Ph il l ip s , J . A . , O r l a n d o , P ., D a v ie s , K . E ., O l d , J . & A n t o n a r a k is , S . E. (1986). Carrier testing in haemophilia A . L ancet i, 148.

J a y e , M., d e l a S a l l e , H., S c h a m b e r , F . , B a l l a n d , A., K o h l i , V., F in d e l i , A., T o ls t o s h e v , P. & L e c o c q , J .-P . (1983). Isolation of a human anti-haemophilic factor IX cDNA clone using a unique 52-base synthetic oligonucleotide probe deduced from the amino acid sequence of bovine factor IX. Nucl. Acids Res. 11 , 2325—2335.

K a ta y a m a , K . , E r ic s s o n , L. H., E n f ie ld , D . L., W a ls h , K . A., N e u r a t h , H., D a v ie , E . W . & TlTANI, K. (1979). Comparison of amino acid sequence of bovine coagulation factor IX (Christmas factor) with that of other vitamin K-dependent plasma proteins. Proc. natn. Acad. S ci. U.SA. 76, 4990-4994.

K u r a c h i , K . & D a v ie , E . W . (1982). Isolation and characterization of a cDNA coding for human factor IX. Proc. natn. Acad. Sci. U.SA. 79, 6461-6464.

M c K e e , P . A . (983). Haemostasis and disorders o f blood coagulation. In The M etabolic Basis o f In h erited D isease, 5th edn (ed. J. B. Stanbury, J. B. Wyngaarden, D. S. Fredrickson, J. L. Goldstein & M . S. Brown), pp. 1531-1560. New York: McGraw-Hill.

M ib a s h a n , R . S., G ia n n e l l i, F., P em b rey , M . E. & R o d e c k , C. H. (1986). The antenatal diagnosis of clotting disorders. In A dvanced M edicine, vol. 20 (ed. M . J. Brown). London: Pitman (in press).

N o y e s , C. M., G r i f f i t h , M. J ., R o b e r ts , H. R . & L u n d b la d , R . L . (1983). Identification of the molecular defect in factor IX (Chapel H ill): Substitution of histidine for arginine at position 145. Proc. natn. Acad. Sci. U.SA. 80, 4200-4202.

O r r , E., W o z n e y , J ., B u e c k e r , J., P it tm a n , D., K a u fm a n , R. & T o o l e , J. (1985). “S p a c e r” fu n c tio n im p lied fo r th e h e a v ily g lyco sy la ted reg ion o f fa c to r VIII. Thromb. H aemostas. 54, 54.

P eake , I. R ., F urlong, B. L . & Bloom, A . L . (1984). C arrier detection by gene analysis in a fam ily w ith haem ophilia B (factor IX deficiency). L ancet i, 242.

R e e s , D. J. G., R iz z a , C. R . & B r o w n le e , G. G. (1985). Haemophilia B caused by a point mutation in a donor splice junction of the human factor IX gene. Nature, Lond. 316, 643-645.

T o o l e , J. J ., K n o p f , J. L ., W o s n e y , J. M ., S u l t z m a n , L. A ., B u e c k e r , J. L ., P it tm a n , D . D ., K a u f m a n , R. J ., B r o w n , E., S h o e m a k e r , C ., O r r , E. C ., A m p h le t t , G. W ., F o s t e r , B . W ., C o u , M . L., K n u t s o n , G. J., F a s s , D . N . & H e w ic k , R. M . (1984). M o lec u la r c lo n in g o f a c D N A en co d in g h u m an an ti-h aem o p h ilic fa c to r. Nature, Lond. 312, 342-347.

W e a t h e r a l l , D. J. (1982). The New G enetics and C linical P ra ctice. London: Nuffield Provincial Hospitals Trust.

W e a t h e r a l l , D. J., H ig g s, D. H ., W o o d , W . G. & C le g g , J. B. (1984). Genetic disorders of human haemoglobin as models for analysing gene regulation. Phil. Trans. R. Soc. Lond. B 307, 247-259.

W in s h ip , P. R ., A n s o n , D. S . , R iz z a , C. R . & B r o w n le e , G. G. (1984). Carrier detection in haemophilia B using two further intragenic restriction fragment length polymorphisms. Nucl. Acids Res. 12, 8861-8872.

458 G. G. Brownlee

W o o d , W . I . , C a p o n , D . J ., S im o n se n , C . C ., Ea t o n , D . L . , G itsch ier , J., K eyt , B ., S e e b u r g , P. H ., S m ith , D. H ., H o l l in g sh e a d , P., W io n , K . L . , D e l w a r t , E ., T u d d e n h a m , E . G . D., V e h a r , G . A . & L a w n , R. M. (1984). Expression of active human factor V I I I from recombinant D N A clones. N ature, Lond. 312, 330-337.

Y o sh ttake , S . , S c h a c h , B. G., F o ster , D . C., D a v ie , E. W. & K u r a c h i, K . (1985). Nucleotide sequence of the gene for human factor IX (antihaemophilic factor B). B iochem istry 24, 3736.