- Top of page
- Patients and methods
The present series comprises all families (n = 77) with haemophilia B in Sweden and may be considered to be representative for the purposes of a population-based study of mutational heterogeneity. The 77 families (38 severe, 10 moderate, 29 mild) had 51 different mutations in total. Thirteen families had total, partial or small deletions, two had mutations in the promoter, eight families had splice site mutations, 14 had nonsense and the remaining 41 had missense mutations. Ten of the mutations, all CT or GA, recurred in 1–6 other families. Using haplotype analysis of seven polymorphisms in the factor IX (FIX) gene, we found that the 77 families carried 65 unique, independent mutations. Of the 48 families with severe or moderate haemophilia, 23 (48%) had a sporadic case of haemophilia compared with 31 families out of 78 (40%) in the whole series. Five of those 23 sporadic cases carried de novo mutations, 11 out of 23 of the mothers were proven carriers and, in the remaining seven families, it was not possible to determine carriership. Eleven of the 48 patients (23%) with severe haemophilia B developed inhibitors and all of them had deletions or nonsense mutations. Thus, 11 out of 37 (30%) patients with severe haemophilia B as a result of deletion/nonsense mutations developed inhibitors compared with 0 out of 11 patients with missense mutations. The ratio of male to female mutation rates was 5·3 and the overall mutation rate was 5·4 × 10−6 per gamete per generation.
Haemophilia B is an X-linked hereditary disorder affecting half the sons of carrier females. The disease is caused by deficient or defective coagulation factor IX (FIX) and, depending on the plasma concentration of FIX clotting activity (FIXC), haemophilia B is classified as severe (FIXC < 1 U/dl), moderate (FIXC 1–4 U/dl) or mild (FIXC 5–25 U/dl). The FIX gene is located in the distal part of the long arm of the X-chromosome (Xq27.1). Yoshitake et al (1985) elucidated the entire nucleotide sequence, which is 33·5 kb long, has eight exons (a–h) and manifests strong homology with other vitamin K-dependent coagulation factors. The Factor IX protein is a serine protease synthesized in the liver via a vitamin K-dependent process and occurs in plasma as a single glycoprotein of 415 amino acids with a molecular weight of 57 000 (Di Scipio et al, 1977).
Haemophilia B is caused by a wide variation of mutations distributed over the entire FIX gene. Since 1990, an annually updated database has been published of characterized point mutations, small deletions and insertions (Giannelli et al, 1998). However, patients with complete or partial deletions or more complex rearrangements are excluded (Thompson, 1991). In the latest update of the database, a total of 1713 mutations are listed, of which 652 are unique molecular events, while the remainder are repeats (Giannelli et al, 1998). It is not clear how many of the repeats are unrelated events and how many may be the result of a founder effect. Most patients with severe or moderate haemophilia and identical mutations may have unique mutations, whereas many patients with mild haemophilia may have mutations with a common derivation (Ketterling et al, 1991). CpG dinucleotides have been shown to be hot-spots for mutations and account for many of the identical mutations found in unrelated families (Cooper & Krawczak, 1990; Green et al, 1990).
The Swedish haemophilia B families have been carefully registered at the haemophilia centres for many decades, during which time pedigrees have been drawn up and maintained. The present series comprises all families with haemophilia B in Sweden and may be considered to be representative for the purposes of a population-based study of mutational heterogeneity. The aim of this study was to use this complete national database of mutations to study the type of mutations, to elucidate whether families with identical mutations share a common ancestor, to define the frequency of mutation and sporadic cases and, finally, to study the correlation between type of mutation and inhibitor development.
- Top of page
- Patients and methods
Table I shows the mutations found in families with severe haemophilia B (n = 38). Twelve out of 38 families had deletions [four total, two partial and six small deletions (< 10 bp), 14 nonsense mutations creating stop codons, nine missense mutations, two splicing defects and two mutations in the promoter]. The index case in family Malmö 7, with sporadic haemophilia, was found to have two mutations, one neutral and one nonsense (Montandon et al, 1990). Four mutations have not previously been reported from other centres to the International Haemophilia B database (Giannelli et al, 1998) (C 6693 T; Gln stop and T 31306 G; Tyr 395 stop and G 10391 C; acceptor splice site and G 17668 C; acceptor splice site). Twenty of the 27 point mutations affected CpG dinucleotides (CT or GA transitions). Two mutations, G −6 C and A 13 G, were in the promoter and caused the Leyden phenotype (Veltkamp et al, 1970). Two families with missense mutations had measurable but low FIXAg levels.
Table I. Mutations in families with severe haemophilia B (n = 38).
|Mö-ID||Nucleotide position||Amino acid change||FIXC||FIXAg|
|53||exon a-f del||–||0||0|
|82||exon a–d, f–h del||–||0||0|
|8||6398 del 2 bp (Fs)||Glu 8 stop 14||0||0|
|9||6466 del 1 bp (Fs)||Val 31, stop 57||0||0|
|10||6665 del 10 bp||splicing defect||0||0·2|
|44||20398 del 1 bp (Fs)||Thr 140 stop 156||0||0|
|13||20510 del 1 bp (Fs)||Asp 177 stop 198||0||0|
|1||30950 del 8 bp (Fs)||Asp 276, stop 288||0||0|
|79||A 13 G||promoter||0||0|
|6||C 6364 T||Arg −4 Trp||0||26|
|4||C 6460 T||Arg 29 stop||0||0|
|69||C 6460 T||Arg 29 stop||0||0|
|52||C 6693 T||Gln 44 stop||0||0|
|83||G 10391 C||acceptor splice||0||0|
|11||G 17668 C||splicing defect||0||0|
|51||C 17700 G||Cys 95 Trp||0||0|
|7||C 17761 T||Arg 116 stop||0||0|
|71||C 17761 T||Arg 116 stop||0||0|
|12||G 20375 T||Cys 132 Phe||0||0|
|5||G 20561 A||Trp 194 stop||0||0|
|61||G 20561 A||Trp 194 stop||0||0|
|3||C 30863 T||Arg 248 stop||0||0|
|14||C 30863 T||Arg 248 stop||0||0|
|15||C 30863 T||Arg 248 stop||0||0|
|41||C 30875 T||Arg 252 stop||0||6|
|65||C 30875 T||Arg 252 stop||0||0|
|7||C 30890 T*||His 257 Tyr*||0||0|
|50||A 30927 T||Asp 269 Val||0||0|
|18||G 31035 A||Gly 305 Asp||0||0|
|58||C 31118 T||Arg 333 stop||0||0|
|24||G 31128 A||Cys 336 Tyr||0||0|
|47||G 31170 A||Cys 350 Tyr||0||0|
|49||G 31276 A||Trp 385 stop||0||0|
|73||T 31306 G||Tyr 395 Trp||0||0|
Table II shows the mutations found in the families with moderate haemophilia B. In one of the families, the only affected member was a woman with deletion of at least part of one of the FIX genes concomitant with extreme lyonization (Kling et al, 1991). The remaining nine patients had point mutations, five with the same mutation (C 6364 T; Arg −4 Trp) and eight mutations affecting CpG dinucleotides. One of the mutations (C 30114 C; His 221 Tyr) is in the proposed active site and had not been described before from other centres.
Table II. Mutations in families with moderate haemophilia B (n = 10).
|Mö-ID||Nucleotide change||Protein change||FIXC||FIXAg|
|33||G 122 A||splicing defect||3||2|
|19||C 6364 T||Arg −4 Trp||3||36|
|20||C 6364 T||Arg −4 Trp||2||27|
|40||C 6364 T||Arg −4 Trp||1||30|
|48||C 6364 T||Arg −4 Trp||2|| |
|75||C 6364 T||Arg −4 Trp||2|| |
|66||C 20413 T||Arg 145 Cys||3|| |
|76||C 30114 T||His 221 Tyr||1|| |
|26||T 31041 G||Val 307 Gly||3||4|
|77||T 31274 C||Trp385 Arg||3|| |
Table III shows the mutations found in families with mild haemophilia B. All patients had point mutations creating amino acid changes, except five families who had the same ‘silent’ mutation (G 17736 A; Val 107, silent) that had not been described before from other centres. Four other mutations had not been described by other centres before (G 10395 T; Gly 48 Val and C 10415 T; Pro 55 Ser and C 10416 T; Pro 55 Leu and C 31248 A; Thr 376 Asn). Twenty-two of the 29 mutations affected CpG dinucleotides.
Table III. Mutations in families with mild haemophilia B (n = 29).
|Mö-ID||Nucleotide change||Protein change||FIXC||FIXAg|
|27||G 10395 T||Gly 48 Val||19||108|
|21||C 10415 T||Pro 55 Ser||12||52|
|22||C 10416 T||Pro 55 Leu||20|| |
|67||T 10482 G||Phe 77 Tyr||19||83|
|35||G 17736 A||Val 107, silent||21||14|
|37||G 17736 A||Val 107, silent||15||24|
|42||G 17736 A||Val 107, silent||20||24|
|55||G 17736 A||Val 107, silent||18|| |
|56||G 17736 A||Val 107, silent||17|| |
|32||G 20414 A||Arg 145 His||6||115|
|36||G 20414 A||Arg 145 His||7||110|
|38||G 20414 A||Arg 145 His||6||86|
|23||G 20414 A||Arg 145 His||7||148|
|17||G 20414 A||Arg 145 His||11||91|
|43||G 20414 A||Arg 145 His||6||115|
|39||T 30100 C||Ile 216 Thr||4||4|
|28||G 30150 A||Ala 233 Thr||22|| |
|29||G 30150 A||Ala 233 Thr||22||12|
|30||G 30150 A||Ala 233 Thr||15|| |
|31||G 30150 A||Ala 233 Thr||11||13|
|46||G 30150 A||Ala 233 Thr||16|| |
|78||G 30864 A||Arg 248 Gln||5|| |
|25||C 31008 T||Thr 296 Met||4||15|
|59||C 31008 T||Thr 296 Met||7|| |
|57||C 31122 A||Ala 334 Asp||27|| |
|62||C 31122 A||Ala 334 Asp||23|| |
|72||G 30150 A||Ala 233 Thr||9|| |
|16||C 31248 A||Thr 376 Asn||15||13|
Families with the same mutation were tested to see whether they were haploidentical, using seven different polymorphisms in the FIX gene. Different haplotypes for one or more polymorphisms were found in two out of two families with C 17761 T (one of the mothers was a non-carrier and only one of the families carried the neutral mutation C 30890 T), two out of two families with G 20561 A, three out of three families with C 30863 T (all three mothers were non-carriers), two out of two families with C 30875 T (the mothers were non-carriers), five out of five families with C 6364 T, one out of five families with G 17736 A, one out of six families with G 20414 A, and one out of five families with G 30150 A. Two families had C 6460 T and were unrelated, owing to different ethnic origin. Malmö 57 and 62 had the same mutation with identical alleles for TaqI, MnlI and HhaI. Malmö 25 and 59 had the same mutation with identical alleles for TaqI, MnlI and HhaI.
Estimation of sex-specific mutation rate
Of the 77 families, 31 had a sporadic case of haemophilia. If one considers only the 48 families with severe or moderate haemophilia (to avoid bias from possible undiagnosed mild haemophiliacs), 23 (48%) had a sporadic case of haemophilia. Five of these 23 sporadic cases carried de novo mutations as the mother did not carry the mutation (and no evidence of mosaicism could be found), 11 out of 23 of the mothers were proven carriers and in the remaining seven cases it was not possible to determine carriership [owing to deletion (three cases) and difficulties in obtaining samples (four cases)]. Five patients carried proven de novo mutations in the 16 families with sporadic mutations in which it was possible to perform a thorough investigation of all relevant family members. Thirty-one families had sporadic mutations and the number of de novo mutants in the whole Swedish male population (4·4 million) may be calculated as 5/16 × 31 = 9·7. The remaining mothers of the sporadic cases in the 16 investigated families carried de novo mutations, according to restriction fragment length polymorphism (RFLP) analysis. By extrapolation to the whole population, this implied that 11/16 × 31 = 21·3 women are new mutants carrying a FIX gene mutation. However, the probability of detecting a new heterozygous mutant through an affected offspring is equal to the risk that the carrier has given birth to a boy with haemophilia. This can be approximately calculated as P = S/4 × (1–C) [1·8/4 × (1–0·24) = 0·34], in which S is the average number of children born to Swedish women during the relevant period and C the fraction of Swedish women below childbearing age (< approximately 20 years) (figures for S and C from the National Bureau of Statistics). Thus, the total number of new female carriers of a FIX mutation in the population should be 21·3/0·34 = 62·6.
Males who are de novo mutants provide a direct estimate of the female-specific mutation rate (u): 9·7/4·4 × 106 = 2·2 × 10−6 (the male population of Sweden is estimated to be 4·4 million). In carrier women who are new mutants, the mutation originates from the maternal grandmother or grandfather. The frequency of such women, 62·6/4·5 × 106 = 13·9 × 10−6, is an estimate of u + v, in which v is the mutation rate in males (the female population of Sweden estimated to be 4·5 million). Thus, v = (13·9–2·2) × 10−6 = 11·7 and the ratio of male to female mutation rates (v/u) = 11·7/2·2 = 5·3. The overall FIX gene mutation rate according to the formula µ = (v + 2 u)/3 is 5·4 × 10−6 per gamete per generation.
Inhibitor development and type of mutation
Eleven of the 48 patients (23%) with severe haemophilia B developed inhibitors and all of them had deletions or nonsense mutations (Table IV). Thus, 11 out of 37 (30%) patients with severe haemophilia B as a result of deletion/nonsense mutations developed inhibitors compared with 0 out of 11 with missense mutations.
Table IV. Mutations in patients who developed inhibitors (n = 11).
|82||del exons A–D, F–H||–|
|44||20398 del 1 bp (fs)||fs Thr 140, stop 156|
|44||20398 del 1 bp (fs)||fs Thr 140, stop 156|
|1||30950 del 8 bp (fs)||fs Asp 276, stop 288|
|4||C 6460 T||Arg 29, stop|
|69||C 6460 T||Arg 29 - stop|
|5||G 20561 A||Trp 194, stop|
|3||C 30863 T||Arg 248, stop|
- Top of page
- Patients and methods
Owing to the careful registration for many decades of all cases of haemophilia in a national registry, the present series of haemophilia B families represents the entire population of Sweden. It was found that almost all families with haemophilia B had a unique mutation, as 84·4% (65 out of 77) of the mutations found were seen to be independent events when families with the same mutations were studied for haplotype identity in the FIX gene. This is a much higher figure than the one found in the latest international database of point mutations, deletions and insertions (Giannelli et al, 1998), in which only 38% were unique events (652 out of 1713), bearing in mind that large deletions were not included. If one considers only the severe and moderate families in our series, all families were unrelated and had unique mutations. Sharing the same haplotype does not prove that the families are unrelated. However, the results were repeated by testing several polymorphisms (avoiding linkage disequilibrium) in many families and, in some families that carried the same mutation, the results were confirmed because the origin of mutation or ethnic origin ruled out the fact that they were related.
Considering only the 66 unique mutations (in 65 families), 53% (35) were missense mutations, 21% (14) were nonsense/stop mutations, 20% (13) were deletions (total or partial), 3% (two) were splice site mutations and 3% (two) were mutations in the promoter (Leyden variants). Seventy-four percent of the mutations (40 of the 54 unique point mutations) were either CT or GA in our series, compared with 59% in the International database (Giannelli et al, 1998) and 50% in the Seattle series (Chen et al, 1991). For example, in the eighth edition of the database (Giannelli et al, 1998), 87 out of 1713 had the C 31008 T compared with only 2 out of 77 in our series. In the series reported by Attali et al (1999), G 20414 A (Arg 145 His) and G 30150 A (Ala 233 Thr) constituted 30% of the 70 unrelated cases. In our series, these mutations were found in 12 out of 29 families, but only 4 out of 12 could be proved as being unique by haplotype analysis. Certainly hot-spots for mutations do exist, but it could well be that certain types of mutations are more frequent in some ethnic groups.
An important issue in genetic counselling is calculating the risk that a mother of a sporadic case of haemophilia is a carrier. In an early report of the origin of mutations in sporadic cases of haemophilia B, we found that in 6 out of 10 carrier mothers of a sporadic case, the mutation had arisen in the healthy maternal grandfather (Kling et al, 1992), thus implying a higher frequency of mutations in the male than the female X-chromosome. In a subsample of the Swedish population, Montandon et al (1992) estimated the male to female mutation rate, v/u, to be 11·0. In a recent report, Green et al (1999) have studied a large subsample (n = 424 families) of the British population and estimated the male to female mutation rate to be 8·64, the female-specific rate 2·18 and the male-specific rate 18·82. Our figure, v/u = 5·3, is in good agreement with the British figure 8·64. Our figures for the female mutation rate are very similar, 2·18 versus 2·2, but our figure for the male mutation rate is slightly lower at 11·7 versus 18·82. Our estimate is also in good agreement with another estimate of overall male to female mutation ratio of 3·75, based on 59 families with known germline origin (Ketterling et al, 1999). In the latter study, it was found that the sex-specific mutation rate differed according to the mutation type, a finding not seen in the British series. Our results are in agreement with the British results: out of five female mutations, four were CT and one was a total deletion, and, in seven proven male mutations, four were CT, one was a 2 bp del, one was a splice and one a GC. In our first report we found paternal age at birth (41·5 years) of a new female carrier to exceed the average paternal age in the general population, a finding that is still relevant in the present series and contradictory to the results of Ketterling et al (1999), who did not find this association between mutation and paternal age but rather a maternal age effect. A higher mutation rate in males has also been shown in haemophilia A (Becker et al, 1996).
The overall rate of FIX gene mutation per gamete per generation that we found in the Swedish population, 6·4 × 10−6, is in excellent agreement with the 7·7 × x 10−6 found by Green et al (1999) and with an old estimate of 3·1 × 10−6 made without the use of molecular biology by Ferrari & Rizza (1986) using a method by Haldane (1935).
In our series, 23% of patients with severe haemophilia B (11 out of 48) developed inhibitors, which is a high figure (Shapiro, 1979; Sultan, 1992). No patients with moderate or mild haemophilia B developed antibodies. The type of mutation is, as reported earlier (Giannelli et al, 1983; Ljung, 1995; Gill, 1999), a distinct predisposing factor for inhibitor development. In the present series, none of the 11 patients with missense mutations developed inhibitors while 11 out of 37 (30%) of the patients with deletions, nonsense or stop codon mutations developed inhibitors. However, in one family with total deletion, none of three affected males developed inhibitors (Wadelius et al, 1988). One may speculate on whether the high incidence of inhibitors in haemophilia B in Sweden is accidental, owing to the limited number of cases, or is real and as a result of, for example, the choice of concentrate or mode of delivery. One important factor is probably the fact that the frequency of large deletions (> 30 bp) seems to be higher in Sweden (8% of the families, but 13% of the total number of patients with severe haemophilia B) than elsewhere (2–5% of reported cases) (Thompson, 1991; Ketterling et al, 1994; Saad et al, 1994), bearing in mind that we do not have reliable data on the frequency of large deletions as these are not included in the International database. In the British series, large deletions were only found in 7 out of 412 families (1·7%) (Green et al, 1999). It is reasonable to assume population-specific differences in the frequency of various types of mutations that may have clinical implications for that population. A similar finding was recently reported in a small series of sporadic cases in Mexico (Jaloma-Cruz et al, 2000).
In summary, in this study of a whole population, a wide spectrum of mutations were found with unique mutations in all families with the severe and moderate forms, a very high proportion of mutations in CpG dinucleotides, and a higher than expected prevalence of deletions. Twenty-one percent of sporadic cases were de novo mutants, the overall frequency of mutations was 5·4 × 10−6 per gamete per generation, and the ratio of male to female mutation rate was 5·3. The type of mutation was a strong predictor of the risk for the development of inhibitors. This national database will be of great value for the genetic counselling of present and future generations in Sweden.