Coinheritance of Factor V (FV) Leiden enhances thrombin formation and is associated with a mild bleeding phenotype in patients homozygous for the FVII 9726+5G>A (FVII Lazio) mutation

Coagulation factor VII (FVII) is a vitamin K-dependent glycoprotein that plays a key role in the initiation of coagulation.¹  Following vascular damage, membrane-bound tissue factor (TF) forms a Ca²⁺-dependent complex either with FVII (which is then converted to FVIIa by a single proteolytic cleavage) or directly with circulating FVIIa, present in blood at a very low concentration. TF-bound FVIIa activates factor X (FX) to FXa, which, together with its cofactor factor Va (FVa), converts prothrombin to thrombin. In addition, the TF/FVIIa complex can also activate factor IX to FIXa, which, after forming a complex with factor VIIIa (FVIIIa), contributes to FXa generation and thereby to thrombin formation via the intrinsic pathway. In plasma, the activity of the TF/FVIIa complex is down-regulated by the tissue factor pathway inhibitor (TFPI), which acts via the formation of a quaternary complex with FXa, FVIIa, and TF.²  Thrombin generation is ultimately shut down by activated protein C (APC), which proteolytically inactivates FVa and FVIIIa, the essential cofactors of the prothrombin- and intrinsic FX-activating complexes.³ 

Severe FVII deficiency^4,5  affects about 1 in 500 000 individuals in the general population and is inherited as an autosomal recessive trait with variable penetrance. Severely affected patients may develop life-threatening hemorrhages and require substitutive treatment with plasma concentrates or recombinant FVIIa.⁶  Several intragenic mutations that impair gene expression or protein secretion (CRM^- deficiency) or alter protein function (CRM⁺ deficiency) have been described (FVII Mutation Database, http://europium.csc.mrc.ac.uk). Moreover, a few intragenic polymorphisms^7-10  that modulate plasma levels of FVII^8,10-12  are known.

The FVII gene is located on chromosome 13 (13q34-qter), next to the FX gene,^13,14  and comprises 9 exons and 8 introns.¹⁵  A polymorphic minisatellite, consisting of a variable number of 37-nucleotide tandem repeats, spans the 3′ end of exon 7 and the 5′ portion of intron 7,⁷  and the first repeat contains the donor splice site for the excision of intron 7. A point mutation (9726+5G>A, FVII Lazio) affecting nucleotide +5 of intron 7, which is part of the splicing consensus sequence, has been identified as a common cause of CRM^- FVII deficiency in central Italy.¹⁶  FVII Lazio homozygous individuals have undetectable FVII levels and are usually severe bleeders.¹⁶  In vitro expression experiments have shown that the FVII Lazio mutation activates a cryptic splice site in the second minisatellite repeat, leading to a 37-nucleotide insertion in the mature transcript, which in turn causes a frameshift and premature termination of translation. Only 0.2% to 1% of all FVII Lazio mRNA is spliced correctly, resulting in a very small amount of normal FVII.¹⁷ 

The poor correlation between FVII levels and bleeding tendency^4,5,18  in carriers of FVII defects points at the existence of additional genetic and/or environmental factors that modulate the clinical phenotype. In particular, coinheritance of common thrombophilic mutations might result in a milder clinical phenotype in patients with severe FVII deficiency, as it has been shown for hemophilic disorders.^19-23  The FVII Lazio mutation, which is relatively frequent in central Italy and predicts a severe bleeding diathesis in the homozygous condition, provides an excellent model to investigate the effects of thrombophilic mutations on severe FVII deficiency. Here we report on the identification of an asymptomatic FV Leiden carrier among 7 patients homozygous for the FVII Lazio mutation. The amounts of FXa, FVa, and thrombin generated after extrinsic activation of plasma in the absence and presence of APC were compared between 2 FVII Lazio homozygotes, the FV Leiden carrier, and a noncarrier. The effect of the FV Leiden mutation on thrombin generation was further investigated in a plasma model of FVII deficiency.

There were 7 patients with severe FVII deficiency due to homozygosity for the FVII Lazio mutation recruited from 2 Italian hemophilia centers. Their demographic and clinical characteristics are summarized in Table 1. Genetic analysis (ie, FVII-FX gene haplotype analysis and the screening for thrombophilic mutations) was performed in all 7 patients. Functional assays (ie, measurement of FXa, FVa, and thrombin generation in plasma in which coagulation was triggered via the extrinsic pathway) were performed only in the patients for whom plasma was available, namely patients no. 3 and no. 6. Informed consent for participation in the study was obtained from all patients.

Venous blood was drawn by venipuncture in 129 mM sodium citrate (9:1 vol/vol). Platelet-poor plasma for functional assays was obtained after 2 centrifugation steps. Whole blood was first centrifuged at 3000g for 25 minutes at room temperature, then the supernatant was transferred to a fresh tube and centrifuged again at 20 000g for 30 minutes at 4°C. Plasma was aliquoted, snap-frozen, and stored at -80°C.

FVII activity (FVII:C) was measured via a one-stage clotting assay in FVII-deficient plasma, using human thromboplastin (Thromborel S; DADE-Behring, Marburg, Germany) as a trigger. FVII antigen (FVII:Ag) was measured with a commercially available enzyme-linked immunosorbent assay (ELISA) kit (Asserachrom VII:Ag; Diagnostica Stago, Asnières, France).

Genomic DNA was extracted from peripheral blood leukocytes according to a standard procedure. The FVII Lazio mutation was detected by polymerase chain reaction (PCR)-mediated amplification of FVII exon 7 (including splicing junctions) followed by RsaI restriction analysis, as described.¹⁶  Genotyping for the FVII and FX gene polymorphisms included in the multipoint haplotype analysis (ie, -402G>A, -401G>T, 5′F7, IVS7 VNTR, and R353Q in the FVII gene, and 5′F10 in the FX gene) was performed as previously described.^10,11,24 

Carriership of the thrombophilic mutations FV R506Q (FV Leiden), FV H1299R (FV R2), and prothrombin (PT) 20210G>A was ascertained by DNA amplification followed by restriction analysis, as reported.^25-27 

Platelet-poor plasma was defibrinated with 1 U/mL Ancrod (WHO International Laboratory for Biological Standards, Hertfordshire, United Kingdom) and the clot was removed with a plastic spatula. After prewarming the plasma at 37°C, the coagulation cascade was triggered with a start mix containing recombinant TF (Dade Innovin; DADE-Behring), synthetic phospholipid vesicles (DOPS/DOPC/DOPE [1,2-dioleoyl-sn-glycero-3-phosphoserine/1,2-dioeoyl-sn-glycero-3-phosphocholine/1,2-dioleoyl-snglycero-3-phosphoethanolamine], 20/60/20, mol/mol/mol), and CaCl₂ (final concentrations in plasma: approximately 6.4 ng/mL TF [approximately 136 pM], 15 μM phospholipids, and 16 mM CaCl₂), in the absence or in the presence of APC (0.4 nM). At regular time points samples were drawn from the plasma mixture and diluted in a “quenching” buffer (50 mM Tris [tris(hydroxymethyl)aminomethane], pH 7.9, 175 mM NaCl, 20 mM EDTA [ethylenediaminetetraacetic acid], 0.5 mg/mL ovalbumin for samples taken for thrombin and FXa quantification; 25 mM HEPES [N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, pH 7.7], 175 mM NaCl, 5 mg/mL bovine serum albumin [BSA], 3 mM CaCl₂, and 1 μM Pefabloc TH for the sample taken for FVa quantification). The latter quenching buffer contained the reversible thrombin inhibitor Pefabloc TH (Pentapharm, Basel, Switzerland) to prevent further activation of FV. Thrombin concentration was measured directly using the chromogenic substrate S2238. FXa and FVa concentrations were quantified via prothrombinase-based assays using purified proteins and synthetic phospholipid vesicles. Reaction conditions were as follows: 40 μM phospholipid vesicles (DOPS/DOPC, 10/90, mol/mol), 5 nM bovine FVa, 0.5 μM human prothrombin, and approximately 3 mM CaCl₂ for the FXa assay; 40 μM phospholipid vesicles (DOPS/DOPC, 10/90, mol/mol), 0.52 nM bovine FXa, 0.5 μM human prothrombin, approximately 3 mM CaCl₂, and 1 μM Pefabloc TH for the FVa assay. Pefabloc TH was added to the FVa assay mixture to prevent further activation of FV by the thrombin generated in the assay. From the amount of thrombin formed (quantified with the chromogenic substrate S2238), the FXa or FVa concentration was calculated using a calibration curve made with known amounts of FXa or FVa. The reaction mixtures used for the construction of the calibration curve for FVa contained 1 μM Pefabloc TH, thereby automatically correcting for the (small) effects of Pefabloc TH on the activity of the prothrombinase complex and on the quantification of thrombin with S2238. Thrombin, FXa, and FVa generation curves were constructed by plotting concentration as a function of subsampling time. Each time course was measured at least in duplicate. Thrombin generation curves were corrected for α₂M-thrombin according to Hemker et al,²⁸  and the area under the curve (the endogenous thrombin potential, ETP)²⁹  was calculated as a measure of the total amount of thrombin generated in plasma.

Preparation of FV-FVII doubly deficient plasma. Congenitally FVII-deficient plasma (FVII:C < 1%) was purchased from George King Bio-Medical (Overland Park, KS), and 15 μg/mL corn trypsin inhibitor (Kordia, Leiden, The Netherlands) was added to prevent contact activation. FV was removed from the plasma by immunodepletion using a monoclonal antibody (RU-FV3B1) directed against the heavy chain of FVa³⁰  coupled to Sepharose beads (CNBr-activated Sepharose 4B; Amersham Pharmacia Biotech, Uppsala, Sweden). After 4 passages over the RU-FV3B1 column, the residual FV concentration in the plasma was 5% of normal, as determined by a prothrombinase-based FV assay.³¹  Measurement of FVIII levels before and after FV-immunodepletion showed that the plasma had not been significantly diluted during the procedure. Freshly prepared FV-depleted FVII-deficient plasma (referred to as FV-FVII doubly deficient plasma from here onward) was aliquoted, snap-frozen, and stored at -80°C until use.

Reconstitution of the FV-FVII doubly deficient plasma with purified FV and variation of the FVII concentration. To mimic the plasma of the FVII Lazio-homozygous patients, FV-FVII doubly deficient plasma was supplemented with 0.2% FVII (in the form of normal plasma) and reconstituted with purified normal FV or FV Leiden,^32,33  or with a 1:1 mixture of normal FV and FV Leiden, to a final FV concentration of 23 nM. Coagulation was triggered with approximately 6.4 ng/mL TF in the presence and absence of 20 nM APC.

In order to vary the FVII concentration, the FV-FVII doubly deficient plasma was mixed with a congenitally FV-deficient plasma (George King Bio-Medical) in various proportions, ranging from 0% to 10% FV-deficient plasma (ie, 0%-10% FVII). Plasma mixtures were then reconstituted with purified normal FV or FV Leiden. In order to obtain measurable thrombin generation curves at all FVII concentrations both in the absence and presence of APC, the TF concentration was set at approximately 1.9 ng/mL, while the APC concentration in the reaction mixture was 10 nM.

Measurement of thrombin generation using a fluorogenic substrate.Thrombin generation in the reconstituted plasma was measured continuously using a low-affinity fluorogenic substrate for thrombin (Z-Gly-Gly-Arg-AMC; BACHEM, Bubendorf, Switzerland). Plasma was mixed with the fluorogenic substrate (300 μM final concentration) in the well of a microtiter plate and prewarmed at 37°C for 3 minutes. Coagulation was triggered with a mixture containing TF, phospholipids, and CaCl₂ in the presence or absence of APC (as described in “Measurement of thrombin, FXa, and FVa generation in extrinsically triggered plasma”) and thrombin generation was followed in time in a fluorometer (SPECTRAmax GEMINI XS; Molecular Devices, Sunnyvale, CA). Raw data from the fluorometer (expressed as relative fluorescence units, RFUs) were corrected for inner-filter effects and substrate consumption and subsequently converted to thrombin concentrations using a thrombin standard kindly provided by Prof H. C. Hemker. The contribution of α₂M-thrombin to the measured signal was calculated and subtracted as described by Hemker et al.³⁴  Corrected data were exported to a Microsoft Excel sheet (Microsoft, Seattle, WA) and averaged with the “moving average” function, after which the first derivative of the data was calculated to obtain the thrombin generation curve.

Of all patients with severe FVII deficiency referred to our lab for mutation screening by 2 hemophilia centers in central Italy, 7 were found to be homozygous for the FVII Lazio mutation (Table 1). Of the patients, 2 (nos. 3-4) were brother and sister, whereas the others were apparently unrelated. All patients had extremely low FVII levels (FVII:Ag ≤ 1%, FVII:C ≤ 2%), and all but one (no. 6) were severe bleeders. FVII deficiency in patient no. 6 was diagnosed at menarche, when she presented with menorrhagia. Since then she has been asymptomatic and had 2 uncomplicated pregnancies without need for treatment.

To verify whether the difference in clinical phenotype between patient no. 6 and the other patients could possibly be attributed to intragenic polymorphisms that modulate FVII levels, a multipoint FVII gene haplotype analysis was carried out in all patients. The following polymorphisms were tested: 2 single-nucleotide polymorphisms¹⁰  and a decanucleotide insertion (5′F7) in the promoter region, a minisatellite in intron 7 (IVS7 VNTR),⁷  and a missense mutation in exon 8 (R353Q).⁸  All FVII Lazio alleles under study turned out to be identical at each polymorphic position (Table 2), in accordance with the founder effect hypothesis for this mutation.¹⁶  Since the FX gene is contiguous to the FVII gene, the haplotype analysis was also extended to a hexanucleotide insertion in the FX gene promoter (5′F10).²⁴  Again, all patients were found to be homozygous for the same allele at this polymorphic site (Table 2).

Screening of the FVII-deficient patients for common thrombophilic mutations, namely FV R506Q (FV Leiden), FV H1299R (FV R2), and PT 20210G>A, revealed that patient no. 6 was heterozygous for FV Leiden, whereas all other patients did not carry any of the mutations.

In order to gain an insight into the molecular mechanism underlying the marked difference in clinical phenotype among patients, plasma was collected from the FV Leiden carrier (patient no. 6) and from a noncarrier (patient no. 3). Coagulation was initiated with TF (approximately 6.4 ng/mL), phospholipids, and CaCl₂, and the generation of FXa, FVa, and thrombin was followed in time by quantifying the amounts of FXa, FVa, and thrombin in samples taken from the plasma mixture at regular time points. To test the influence of FV Leiden, the experiments were performed both in the absence and in the presence of APC (0.4 nM). A normal plasma was included in the analysis as a control.

In the normal control (Figure 1), fast coagulation factor activation was observed. The FXa generation curve reached a peak after 1.5 minutes, and FV was quantitatively activated within 1 minute. This resulted in the rapid generation of a large amount of thrombin. The addition of APC considerably reduced the generation of FVa, but it had no effect on FXa and thrombin generation at the TF concentration used in our experimental set-up.

In the FVII-deficient patients (Figure 2), the activation of all coagulation factors was significantly delayed compared with the normal control. The FXa, FVa, and thrombin generation curves showed a lag phase of several minutes, and considerably lower amounts of FXa and thrombin were formed. FXa formation was affected the most by the lack of FVII. At the peak of the FXa generation curve, about 60 pM FXa (< 1% of the normal control) was formed in both patients. In the absence of APC, more FXa was formed in the plasma of patient no. 6 than in the plasma of patient no. 3 (Figure 2A). The addition of APC (Figure 2D) resulted in further down-regulation of FXa generation in the plasma of patient no. 3, but not in the plasma of patient no. 6, who carried the FV Leiden mutation. FVa generation was quantitatively not affected by the deficiency of FVII. In the absence of APC (Figure 2B) the time courses of FVa generation obtained in the 2 patients were superimposable, whereas in the presence of APC (Figure 2E) more FVa was formed in the plasma of patient no. 6 than in that of patient no. 3, which is in line with their respective FV genotypes. As a consequence of the sharp reduction in FXa generation, thrombin formation was also decreased in the FVII-deficient patients compared with the normal control. However, thrombin generation was less dramatically affected than FXa generation. In the absence of APC, the thrombin peak was about 25% (patient no. 3) to 28% (patient no. 6) of that measured in normal plasma (as well as considerably delayed), while the ETP was about 70% of that of normal plasma. Between the 2 patients there was hardly any difference in the amount of thrombin formed in the absence of APC (Figure 2C). However, in the presence of APC (Figure 2F) thrombin generation in patient no. 3 showed a slightly longer lag phase and the total amount of thrombin generated was about half (56%) of that of patient no. 6.

Since the difference in thrombin generation between the 2 patient plasmas might theoretically be due to other plasma variables in addition to the FV, we decided to investigate the effect of the FV Leiden mutation on thrombin generation in a model system of FVII deficiency, that is, FV-FVII doubly deficient plasma supplemented with a variable amount of FVII and reconstituted with purified normal FV or FV Leiden.

To mimic FVII Lazio-homozygous patients, FV-FVII doubly deficient plasma was supplemented with 0.2% FVII¹⁷  and reconstituted with normal FV, FV Leiden, or, to simulate heterozygosity for FV Leiden, a 1:1 mixture of normal FV and FV Leiden. Reconstituted plasmas were triggered with TF (approximately 6.4 ng/mL), phospholipids, and CaCl₂ in the presence and absence of 20 nM APC, and thrombin generation was followed with a thrombin-specific fluorogenic substrate. As shown in Figure 3A, the 3 plasmas yielded superimposable thrombin generation curves in the absence of APC. However, in the presence of APC (Figure 3B), considerably less thrombin (approximately 64%) was formed in the plasma containing normal FV than in the plasmas containing FV Leiden. A normal plasma measured under the same experimental conditions is shown for comparison in Figure 3C.

Comparison of Figure 3 with Figures 1, 2 shows that thrombin generation curves obtained with the fluorogenic method are about 2.4 times higher than the corresponding curves measured with the subsampling method. This is due to the fact that the fluorogenic substrate present in the reaction mixture competes with the physiologic substrates and inhibitors of thrombin in plasma. The predominant effect of this interference is the protection of thrombin from inhibition by antithrombin and other plasma inhibitors, which likely explains the increased levels of thrombin measured in reaction mixtures containing fluorogenic substrate.

To further characterize the interaction between FVII deficiency and FV Leiden, we also investigated the effect of varying FVII levels on thrombin generation in FV-FVII doubly deficient plasma reconstituted with either normal FV or FV Leiden and triggered with approximately 1.9 ng/mL TF (Figure 4). Both in the absence and in the presence of APC (10 nM), thrombin generation progressively increased with increasing FVII concentration. In the absence of APC, thrombin generation curves obtained at each particular FVII concentration were very similar between plasma mixtures reconstituted with normal FV or FV Leiden (data not shown). However, in the presence of APC (Figure 4), plasma reconstituted with FV Leiden (closed symbols) yielded clearly higher and faster thrombin generation than the corresponding plasma reconstituted with normal FV (open symbols). Interestingly, the difference in lag time and peak height of thrombin generation between plasmas containing normal FV and FV Leiden decreased at increasing FVII concentration and became negligible at FVII levels 10% or higher (Figure 4). The same trend was observed for the total amount of thrombin formed, calculated from the area under the thrombin generation curve. The amount of thrombin generated in plasma reconstituted with normal FV was approximately 40% of that formed in plasma reconstituted with FV Leiden at very low FVII (≤ 0.1%), and gradually increased to 86% at 10% FVII. At 50% FVII (data not shown) the total amount of thrombin generated in plasma reconstituted with normal FV was almost the same (95%) as that generated in plasma reconstituted with FV Leiden.

The variable penetrance of genetic defects predicting severe coagulation factor deficiencies indicates that additional genetic and/or environmental variability modulates the clinical expression of these defects. The search for such modifiers has been mainly targeted to hemophilic disorders, due to their relatively high prevalence in the population. In the case of hemophilia A, evidence has been provided that coinheritance of the FV Leiden^19,20  or PT 20210G>A mutations²²  can ameliorate the clinical phenotype,^19,20  and that FV Leiden increases thrombin generation as measured in vitro.³⁵  In addition, the onset of symptoms in children with hemophilia A was found to be significantly delayed in carriers of thrombophilic defects such as FV Leiden, PT 20210G>A, and protein C deficiency.²¹  More recently, a case of severe hemophilia B with mild clinical expression attributable to carriership of FV Leiden has also been reported.²³ 

In contrast to hemophilic disorders, which affect the propagation phase of thrombin generation, FVII deficiency impairs the initiation phase.³⁶  Therefore it is questionable whether thrombophilic mutations that enhance downstream procoagulant reactions can compensate for the inefficient initiation of coagulation due to a FVII defect. The FVII Lazio mutation offers a unique opportunity to investigate the potentially protective role of common thrombophilic mutations in patients with severe FVII deficiency, since (1) its prevalence in the Italian population is compatible with the recruitment of homozygous patients; (2) the very low FVII levels associated with the homozygous state cause a severe but not lethal bleeding diathesis, which favors the detection of other inherited or acquired conditions ameliorating the clinical phenotype; and (3) being a splicing mutation, it results in a very low amount of normal FVII molecules, a condition that enables “mimicking” in a model system. The design of a parallel model system would have been complicated in the case of other mutations causing FVII deficiency through poor secretion of altered molecules or normal amounts of dysfunctional molecules.

Our study was based on the comparison of 7 FVII Lazio homozygous patients who had equally low FVII levels but different clinical outcomes (Table 1). While 6 of them were severely affected, the only FV Leiden carrier of the group had been asymptomatic since the age of 14 years, in spite of repeated challenges of the hemostatic system. This striking difference in clinical presentation could not be attributed to possible up-regulation of FVII levels by other intragenic functional polymorphisms, because an extended haplotype analysis showed that all patients had inherited 2 identical copies of the same FVII allele (Table 2).

The efficiency of the coagulation cascade in the plasma of 2 patients, 1 of whom was the FV Leiden carrier, was evaluated at 3 different levels (FXa, FVa, and thrombin generation) after extrinsic activation of plasma in the absence and presence of APC. All activation steps were significantly delayed in the plasma of the 2 FVII-deficient patients, and considerably less FXa and thrombin were formed than in a normal plasma. Comparing the 2 patients, coagulation factor activation was more efficient in the plasma of the FV Leiden carrier, particularly in the presence of APC. Better discrimination in the presence of APC provides evidence that the observed differences between the 2 patients are indeed attributable to FV Leiden. Moreover, comparison between the thrombin generation curves obtained in the plasmas of the 2 patients (Figure 2C,F) indicates that relatively small differences in the amount of thrombin formed may underlie major differences in clinical phenotype.

The amount of FXa generated in the plasmas of the FVII-deficient patients appeared to be 2 orders of magnitude lower than that of FVa on a molar basis, which indicates that FXa is the limiting component in thrombin formation. The ability of APC to down-regulate FXa generation in the plasma of the FVII-deficient patient without the FV Leiden mutation (Figure 2A,D) was an intriguing finding. A possible explanation for this phenomenon is that FIXa, formed either by the TF/FVIIa complex³⁷  or by the thrombin feedback on FXI,³⁸  contributes to FXa generation via the intrinsic coagulation pathway. A contribution of the intrinsic FX-activating complex would make FXa generation partially dependent on FVIIIa, which is inactivated by APC, leading to reduced FXa generation in the presence of APC. Since FV stimulates APC-mediated FVIIIa inactivation³⁹  and FV Leiden is a poor APC cofactor,^40,41  this hypothesis would also explain the relative inability of APC to decrease FXa formation in the FV Leiden carrier. The fact that more FXa is formed in the plasma of the FV Leiden carrier than in the noncarrier even in the absence of added APC is probably due to the endogenous APC formed by activation of plasma protein C.

In a study on hemophilia A it was observed that not all patients carrying FV Leiden were characterized by a mild clinical course,⁴²  suggesting that FV Leiden alone might not be sufficient to confer protection against bleeding. In addition, FV Leiden and other thrombophilic defects were found to be rare or absent in a group of severe hemophiliacs with a mild bleeding diathesis,⁴³  indicating that the mild phenotype in these patients was due to other (unknown) mitigating factors. To demonstrate that FV Leiden contributes to the increased thrombin generation in patient no. 6, we performed experiments in FV-FVII doubly deficient plasma reconstituted with either normal FV or FV Leiden. Since in this model the FV used for reconstitution is the only variable between the plasma samples, the experiment presented in Figure 3 clearly demonstrates that the presence of FV Leiden can account for the observed differences in thrombin generation in severely FVII-deficient plasmas, at least in the presence of APC. Moreover, variation of the FVII concentration in the reconstituted FV-FVII doubly deficient plasma indicated that the effect of FV Leiden on thrombin generation is most pronounced at extremely low FVII levels and gradually disappears as FVII levels increase (Figure 4). This finding further suggests that the protective effect of FV Leiden in severe FVII deficiency might be restricted to FVII mutations that predict unmeasurable FVII levels. Mutation-specific protection against bleeding by FV Leiden has already been proposed for hemophilia A.⁴² 

Overall, our findings indicate that coinheritance of FV Leiden leads to enhanced thrombin generation and can mitigate the bleeding phenotype in patients with severe FVII deficiency due to homozygosity for the Lazio mutation. In addition, our plasma model of FVII deficiency shows that the ability of FV Leiden to increase thrombin generation is maximal at the lowest FVII concentrations and is lost at higher FVII levels.

The authors wish to thank Prof H. C. Hemker and Dr S. Béguin (Synapse b.v., Maastricht, the Netherlands) for helpful suggestions, as well as Dr T. Lindhout for assistance in the use of the fluorometer.