- Top of page
- Materials and methods
Summary. Previous reports on the effect of Secretor and Lewis blood groups on plasma factor VIII–von Willebrand factor (FVIII–VWF) levels have produced conflicting findings. To determine whether either or both loci can influence plasma FVIII–VWF complex levels, we studied the relationship between Secretor and Lewis genotypes, determined definitively using polymerase chain reaction–restriction fragment length polymorphism analysis, and plasma FVIII coagulant activity (FVIII:C) and VWF antigen (VWF:Ag) levels in 136 healthy volunteers. Overall, significantly higher VWF:Ag levels were found in those individuals homozygous for the Se allele (genotype SeSe) than in those heterozygous for the Se allele (P < 0·001). To minimize any confounding influence of ABO genotype/phenotype, we investigated the relationship between Secretor genotype and plasma FVIII–VWF levels within individuals of the same ABO blood group genotype. In the subgroup analysis of group O1O1 individuals alone, VWF:Ag levels were again significantly higher in those individuals with Secretor genotype SeSe than in those either heterozygous or homozygous for the se null allele. Among A1O1 subjects, homozygous Secretors also had significantly higher VWF:Ag levels. In contrast, we found no relationship between Lewis genotype and either VWF:Ag or FVIII:C levels. This study is the first based on genotypic rather than serological analysis, and resolves the previously confounding effects of the Lewis and Secretor loci on plasma FVIII–VWF complex levels.
Factor VIII (FVIII) and von Willebrand factor (VWF) are plasma glycoproteins that perform vital roles in normal haemostasis (Sadler, 1998). In plasma, FVIII and VWF circulate as a non-covalent complex (Vlot et al, 1998). VWF acts as a carrier molecule for FVIII, protecting it from premature dissociation or degradation (Vlot et al, 1996). The normal population distribution of plasma FVIII–VWF levels shows a wide range, with a skewed distribution towards higher levels (Gill et al, 1987). Plasma FVIII–VWF levels are of clinical significance. Low plasma levels of either FVIII (Haemophilia A) or VWF (von Willebrand disease) have long been recognized as causes of excess bleeding (Nichols & Ginsberg, 1997). Conversely, there is recent evidence that elevated FVIII–VWF levels may represent an important and prevalent risk factor for both ischaemic heart disease (Rumley et al, 1999) and venous thromboembolism (Koster et al, 1995; Kyrle et al, 2000). It is therefore of interest to determine which genes influence plasma FVIII–VWF levels and to understand the mechanism by which these genes act.
It is well established that gene loci other than the FVIII gene (Xq28) and the VWF gene (12p12) can exert major quantitative effects on plasma FVIII–VWF complex levels. The most important of these loci is the ABO blood group locus on chromosome 9q34 (Shima et al, 1995). VWF antigen (VWF:Ag), ristocetin cofactor activity and botrocetin cofactor activity (Gill et al, 1987; Shima et al, 1995), and FVIII coagulant activity (FVIII:C) and FVIII antigen (FVIII:Ag) (McCallum et al, 1983) are both approximately 25% lower in group O individuals than in non-O individuals (groups A, B and AB). Effects of the Secretor blood group locus on plasma concentration of FVIII–VWF have also been reported, albeit inconsistently (Orstavik et al, 1989; Green et al, 1995). Furthermore, a relationship between Lewis blood group phenotype and FVIII–VWF levels has been recently described (Green et al, 1995).
Similar effects from the ABO, Lewis and Secretor blood group systems are plausible because they are closely related (Watkins, 1996). All three systems are characterized by the presence or absence of specific terminal carbohydrate determinants on the oligosaccharide structures of glycoproteins or glycolipids. In the case of Secretor, blood group O (H) and Lewis, the terminal sugar is a fucose. The Secretor blood group locus maps to chromosome 19q13 and has recently been cloned (Kelly et al, 1995). The gene encodes an α(1,2) fucosyltransferase (FUT2) similar to that encoded by the closely related H gene locus (FUT1) necessary for expression of ABO blood groups. Individuals with an Se allele (genotypes SeSe or Sese respectively) are Secretors, and can secrete ABH and Lewis structures in their plasma and secretions. Approximately 20% of Caucasians have the genotype sese, and are non-Secretors (Kukowska-Latallo et al, 1990). The molecular basis underlying polymorphism at the Secretor locus has recently been established. In Caucasians, the se null allele results from an enzyme inactivating point mutation (W143X; Trp 143ter) (Kelly et al, 1995; Vestergaard et al, 1999).
The Lewis blood group maps to chromosome 19p, and encodes an α(1,3) fucosyltransferase (FUT3) which can modify secreted sugar chains (Orntoft et al, 1997). Individuals carrying an Le allele express this glycosyltransferase, while individuals homozygous for le (genotype lele) do not. In Caucasians, two common mutations (T59G and the C314T-T202C haplotype respectively) have been shown to produce an le null allele. C314T and T202C are in complete linkage disequilibrium but virtually never occur on the same allele as the T59G (Elmgren et al, 1996; Orntoft et al, 1997). The molecular genetic mechanism behind non-functional Lewis alleles is not fully understood, but these mutations appear to cause changes in pFUT3 membrane insertion and folding respectively (Elmgren et al, 1996).
Lewis blood group antigens (Lea and Leb respectively) are produced by interaction of FUT2 and FUT3 (Watkins, 1980). Individuals expressing FUT 3 (LeLe or Lele at the Lewis locus) can synthesize Lea structures. Those individuals also expressing FUT 2 (SeSe or Sese at the Secretor locus) can convert Lea into Leb. Consequently, individuals with red cell phenotype Le(a+b–) are non-Secretors, while individuals with phenotype Le(a–b+) are Secretors. Lewis-negative individuals Le(a–b–) may or may not be Secretors.
Previous reports on the effect of Secretor and Lewis blood groups on plasma FVIII–VWF levels have produced conflicting findings, strongly dependent on ABO blood group (Orstavik et al, 1989; Green et al, 1995). However, these studies used serologically determined red blood cell Lewis phenotypes to infer probable Lewis and Secretor genotypes respectively. Recent evidence has shown this approach to be unreliable (Henry et al, 1995; Svensson et al, 2000). Moreover it does not permit discernment of dosage effects. To determine whether either or both loci can influence plasma FVIII–VWF complex levels, we studied the relationship between Secretor and Lewis genotypes, determined definitively using polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP) analysis, and plasma FVIII:C and VWF:Ag levels in 136 healthy volunteers. The validity of this genotyping methodology in European Caucasians has previously been established (Vestergaard et al, 1999).
- Top of page
- Materials and methods
Orstavik et al (1989) were the first to describe an effect of Secretor and Lewis blood groups on plasma concentrations of VWF–FVIII. In a study of 323 twins and 58 blood donors, they found significantly higher VWF:Ag and FVIII:Ag levels in group O individuals with the red cell phenotype Le(a+b–) than in group O individuals with phenotypes Le(a–b+) or Le(a–b–) respectively. However, this relationship was not observed among the non-O (A, B and AB respectively) individuals studied. Moreover, because this was a serologically based study, 80% of those with Le (a–b–) phenotype would have been Secretors, with H substance present in plasma.
Green et al (1995) also reported an association between red cell Lewis phenotype and plasma FVIII–VWF levels. In healthy white men with blood group A, B or AB, both FVIII and VWF:Ag levels were significantly higher in those with Lewis phenotype Le(a–b–) than in Le(a+b–) and Le(a–b+) individuals. This effect of Lewis status on FVIII–VWF levels was not reproduced in group O individuals, women or black men. Moreover, the widely cited conclusion of Orstavik et al (1989) that higher levels of FVIII–VWF are present in group O individuals with Lewis phenotype Le(a+b–) was not verified by this much larger study.
The discrepant findings of these two studies may in part be attributable to differences in the populations studied. However, both groups also determined Lewis and Secretor status by routine serological testing, making it difficult to disentangle the separate effects of the Lewis and Secretor gene products which are obscured in such tests. Moreover, recent reports have demonstrated that such serological determinations can themselves be unreliable, particularly in blood group A1 individuals and in those of non-European origin (Henry et al, 1995; Svensson et al, 2000). Even in European populations, the use of serological Lewis phenotyping to deduce Secretor status has been shown to produce erroneous results in up to 10% cases (Svensson et al, 2000). With the recent description of the molecular bases underlying polymorphisms at both the Lewis and Secretor blood group loci, accurate genotyping is now possible and is widely regarded as definitive (Vestergaard et al, 1999). Not only is this approach free from the above pitfalls of serology, it allows the assessment of dosage effects at the loci, lending power to the analyses. We have therefore used a genotyping strategy to definitively determine whether the Secretor or Lewis loci can exert quantitative effects on plasma FVIII–VWF levels.
Our results clearly demonstrate that genotype at the Secretor blood group locus is a determinant of plasma VWF:Ag concentration. Among group O1O1 subjects, those individuals homozygous for the Se allele (homozygous Secretors) had VWF:Ag levels significantly higher than those individuals with genotypes Sese or sese (P = 0·02 and P = 0·04 respectively). Moreover, in contrast to previous reports linking Secretor status and FVIII–VWF levels (Orstavik et al, 1989), we found this relationship to be also present in A1O1 individuals, and to remain when both A and O were considered as one group. Although FVIII:C levels also demonstrated a trend towards highest levels in SeSe individuals, this association failed to reach statistical significance, suggesting that any effect of Secretor genotype on FVIII levels is probably secondary to a direct effect on VWF:Ag levels.
It is important to note that this effect would not have been detected by a serological analysis because Se heterozygotes and homozygotes would have been grouped together as Secretors. The lack of a dominant Se effect is somewhat surprising and leads us to conclude that it represents a dosage effect such that the products of two functional Se genes are required to influence VWF levels.
The mechanism by which Secretor genotype influences plasma VWF:Ag levels remains unclear. Theoretically, Secretor status may alter the rate of VWF synthesis or secretion within endothelial cells. Alternatively, Secretor status may affect VWF plasma clearance rates. We have previously found a close association between the amount of H substance expressed on circulating VWF and plasma VWF:Ag concentration, suggesting that terminal fucose residues may be important in mediating VWF clearance (unpublished observations). In homozygous Secretors, plasma H substance could theoretically compete with the H determinants expressed on the N-linked glycans of VWF for this clearance mechanism.
It is also possible that that the relationship between Secretor locus and plasma VWF levels is not owing to a direct functional effect of the Secretor gene. Rather the Secretor locus may exist in linkage disequilibrium with another unidentified VWF regulatory locus. The proximity of the Secretor (FUT2) gene to the H (FUT1) gene at 19q raises the possibility that the Se gene is in linkage disequilibrium with a quantitative variant at the latter locus. If this were the case then we would expect the amount of H antigen on VWF in group O individuals to vary between Secretors and non-Secretors. We therefore proceeded to assess this variable but found no difference between the groups.
In contrast to the influence of Secretor genotype on plasma VWF:Ag levels, we found no relationship between Lewis genotype and plasma FVIII–VWF levels. The previously identified association between Le(a–b–) and increased FVIII–VWF was not reproduced. Previous studies may have been confounded by the inability to detect the Secretor status of these individuals. However, the number of individuals with genotype lele in our Caucasian population was predictably low (13/136, 9·6%).
In conclusion, we have shown that individuals homozygous for the functional Se gene have higher plasma levels of VWF and that there is no effect from Lewis genotype. The Secretor effect is seen in both blood group A and O individuals. This study is the first based on genotypic rather than serological analysis and resolves the confounding effects of the Lewis and Secretor loci interaction that may be responsible for the previously conflicting reports of this phenomenon. We postulate that the effect is mediated by competition for clearance mechanisms between H substance on VWF and that in plasma.