Homozygous F5 deep-intronic splicing mutation resulting in severe factor V deficiency and undetectable thrombin generation in platelet-rich plasma

Authors

  • E. CASTOLDI,

    1. Department of Biochemistry, Cardiovascular Research Institute Maastricht (CARIM), Maastricht University, Maastricht, the Netherlands
    Search for more papers by this author
    • These authors contributed equally to this work.

  • C. DUCKERS,

    1. Department of Biochemistry, Cardiovascular Research Institute Maastricht (CARIM), Maastricht University, Maastricht, the Netherlands
    Search for more papers by this author
    • These authors contributed equally to this work.

  • C. RADU,

    1. Department of Cardiologic, Thoracic and Vascular Sciences, 2nd Chair of Internal Medicine, University of Padua Medical School, Padua
    Search for more papers by this author
  • L. SPIEZIA,

    1. Department of Cardiologic, Thoracic and Vascular Sciences, 2nd Chair of Internal Medicine, University of Padua Medical School, Padua
    Search for more papers by this author
  • V. ROSSETTO,

    1. Department of Cardiologic, Thoracic and Vascular Sciences, 2nd Chair of Internal Medicine, University of Padua Medical School, Padua
    Search for more papers by this author
  • G. TAGARIELLO,

    1. Regional Centre for Blood Diseases and Haemophilia, ‘Castelfranco Veneto e Montebelluna’ Hospital, Castelfranco Veneto, Italy
    Search for more papers by this author
  • J. ROSING,

    1. Department of Biochemistry, Cardiovascular Research Institute Maastricht (CARIM), Maastricht University, Maastricht, the Netherlands
    Search for more papers by this author
  • P. SIMIONI

    1. Department of Cardiologic, Thoracic and Vascular Sciences, 2nd Chair of Internal Medicine, University of Padua Medical School, Padua
    Search for more papers by this author

Elisabetta Castoldi, Department of Biochemistry, Maastricht University, PO Box 616, 6200 MD Maastricht, the Netherlands.
Tel.: +31 43 3884160; fax: +31 43 3884159.
E-mail: e.castoldi@maastrichtuniversity.nl

Abstract

Summary.  Background: Coagulation factor (F) V deficiency is associated with a bleeding tendency of variable severity, but phenotype determinants are largely unknown. Recently, we have shown that three patients with undetectable plasma FV and mild bleeding symptoms had sufficient residual platelet FV to support thrombin generation in platelet-rich plasma (PRP). Therefore, we hypothesized that FV-deficient patients with severe bleeding manifestations may lack platelet FV. Objectives: To characterize a FV-deficient patient with a severe bleeding diathesis. Patients/Methods: We performed FV mutation screening and functional studies in a 31-year-old male (FV:C < 1%) with umbilical bleeding at birth, recurrent hemarthrosis and muscle hematomas, and a recent intracranial hemorrhage. Results: The proband was homozygous for a deep-intronic mutation (F5 IVS8 +268A→G) causing the inclusion of a pseudo-exon with an in-frame stop codon in the mature F5 mRNA. Although platelet FV antigen was detectable by immunoprecipitation followed by Western blotting, no FV activity could be demonstrated in the proband’s plasma or platelets with a prothrombinase-based assay. Moreover, no thrombin generation was observed in PRP triggered with 1–50 pm tissue factor (even in the presence of platelet agonists), whereas an acquired FV inhibitor was excluded. Clot formation in the proband’s whole blood, as assessed by thromboelastometry, was markedly delayed but not abolished. Conclusions: This is the first report of a pathogenic deep-intronic mutation in the F5 gene. Our findings indicate that the minimal FV requirement for viability is extremely low and suggest that thrombin generation in PRP may predict bleeding tendency in patients with undetectable plasma FV.

Introduction

Coagulation factor (F) V [1] is an essential clotting factor whose active form (FVa) is required for the efficient conversion of prothrombin into thrombin. FV is synthesized in the liver and is subsequently secreted in the blood stream, where it circulates as an inactive single-chain precursor with an A1-A2-B-A3-C1-C2 domain structure. When coagulation is initiated, FV is activated through limited proteolysis by thrombin or FXa [2]. In this process the B domain is released, yielding a heavy chain (A1-A2, 105 kDa) and a light chain (A3-C1-C2, 71/74 kDa) linked via a Ca2+-ion. FVa acts as a non-enzymatic cofactor of FXa and accelerates FXa-catalysed prothrombin activation more than 1000-fold [3]. FVa activity is down-regulated by activated protein C (APC)-mediated proteolysis of the heavy chain at Arg306, Arg506 and Arg679 [4,5].

Approximately 20% of the FV present in blood is stored in platelet α-granules, from which it is released upon platelet activation [6]. Platelet FV is not synthesized endogenously, but originates from endocytosis and intracellular processing of plasma FV by bone marrow megakaryocytes [7–11]. Platelet FV is stored in a partially activated form and is more resistant to APC-mediated inactivation than plasma FV(a) [9,12].

Congenital FV deficiency (Owren parahemophilia, MIM #227400) is an autosomal recessive bleeding disorder with an estimated incidence of 1:106 [13,14]. The clinical presentation of severe (homozygous) FV deficiency ranges from occasional mucosal bleeding to life-threatening hemorrhages [15–17], without a clear-cut correlation with residual plasma FV levels. To date, > 100 detrimental FV gene (F5) mutations (including missense, nonsense and splicing mutations as well as small insertions/deletions) have been described in FV-deficient patients [18].

While the complete absence of FV is generally considered incompatible with life (based on the uniform embryonal/perinatal lethality of FV knock-out mice [19]), several lines of evidence indicate that the minimal FV requirement for viability is well below 1% [20–22]. In previous studies, we have identified residual platelet FV in combination with low plasma levels of the anticoagulant protein tissue factor pathway inhibitor (TFPI) as a possible rescue mechanism in patients with undetectable plasma FV [23,24]. In particular, we have shown that three patients with undetectable plasma FV and relatively mild bleeding symptoms had sufficient residual FV in their platelets to support thrombin generation in platelet-rich plasma (PRP), which may protect them from major bleeding [24]. Therefore, FV-deficient patients with severe bleeding manifestations might have less or no platelet FV. In the present study we report on a FV-deficient patient with a severe bleeding tendency and undetectable thrombin generation in PRP.

Materials and methods

Blood collection and plasma preparation

Venous blood was obtained from the proband, his parents and four healthy controls on two different occasions with an interval of 8 months. From each donor, 40 mL blood were drawn in 129 mm sodium citrate (1/10 vol/vol) for thrombin generation and thromboelastometry experiments, and 20 mL blood were drawn in 80 mm sodium citrate, 52 mm citric acid and 183 mm glucose (ACD, 1/7 vol/vol) for platelet isolation (see ‘Platelet preparation’) and platelet RNA extraction (see ‘Genetic analysis’). PRP and platelet-poor plasma (PPP) were prepared from citrated plasma as described before [24] and immediately tested in the thrombin generation assay. Buffy coats were stored at −20 °C for later DNA isolation. The study was conducted according to the Helsinki protocol and all subjects provided informed consent to participation.

Thrombin generation

Thrombin generation was measured using the Calibrated Automated Thrombogram method [25], essentially as previously described [24]. Briefly, coagulation in PPP (with 20 μm added phospholipids, TGT-lipids, Rossix, Mölndal, Sweden) and PRP (platelet count standardized to 1.5 × 108 platelets mL−1) was triggered by the addition of 1–50 pm tissue factor (TF; Innovin, Dade-Behring, Marburg, Germany) followed by recalcification (16 mm added CaCl2). In some experiments with PRP, platelets were preactivated with 10 μg mL−1 collagen (Dynamyte Medical, München, Germany) or 20 μm Ca2+-ionophore (A23187; Sigma, Buchs, Switzerland). All measurements were performed in the presence of 32 μg mL−1 corn trypsin inhibitor (Haematologic Technologies, Essex Junction, VT, USA) to prevent contact activation.

To exclude the presence of inhibitory anti-FV antibodies in the proband’s plasma, normal pooled plasma was mixed with the proband’s plasma or with a commercial congenitally FV-deficient plasma (George King Biomedical, Overland Park, KS, USA) to final concentrations of 0%, 0.5%, 1%, 2.5%, 5% and 10%. Plasma mixtures were incubated at room temperature for 2 h and FV activity was subsequently assayed by measuring thrombin generation at 13.6 pm TF.

Platelet preparation

Platelets were isolated and washed as described previously [8]. Washed platelets were divided into two aliquots: one (0.7 × 109 platelets mL−1) was frozen as such, whereas the other (0.5 × 109 platelets mL−1) was activated with thrombin and Ca2+-ionophore as described by Duckers et al. [24] and then frozen.

Platelet lysates, prepared by thawing non-activated platelets in the presence of Triton X100 (Fluka, Buchs, Switzerland) and protease inhibitors [24], were used for the determination of platelet FV antigen levels (ELISA). Activated platelet suspensions were the starting material for platelet FVa immunoprecipitation and activity measurement.

Measurement of FV antigen levels

Plasma and platelet FV antigen levels were measured using an ELISA assay, as described before [24].

Measurement of FV activity levels

Plasma and platelet FV activity levels were determined with a prothrombinase-based assay, as described previously [24]. Assay conditions were as follows: 5 nm FXa, 1 μm prothrombin, 40 μm phospholipid vesicles (DOPS/DOPC, 10/90 mol/mol), 2.5 mm CaCl2, and limiting amounts of FVa.

To investigate to what extent the platelet-dependent stimulation of prothrombin activation was attributable to platelet FVa rather than to other platelet components, the assay was performed in the absence and presence of specific FV inhibitors (anti-FV antibodies, APC/protein S). However, because these inhibitors were unable to completely block platelet FVa activity in control platelets, EDTA was used as an alternative means to abolish FVa activity. Chelation of Ca2+-ions by EDTA results in the dissociation of the heavy and light chains of FVa, causing loss of cofactor activity. At low FVa concentrations this effect is largely irreversible, as only 10% of the original FV activity was recovered when the free Ca2+ concentration was restored. In detail, 11.4 mm EDTA was added to the activated platelet suspension (containing 4.3 mm CaCl2) and incubated for 10 min at room temperature. Subsequently, the original Ca2+ concentration was restored, and platelet FVa activity was assayed immediately. The difference in prothrombinase activity before and after EDTA treatment was taken as a measure of FVa activity.

FVa immunoprecipitation

The FV heavy chain was immunoprecipitated from plasma or activated platelet suspensions with a monoclonal anti-FV heavy chain antibody (3B1, a kind gift from B. N. Bouma) coupled to protein G Sepharose beads (Protein G Sepharose 4 Fast Flow; GE Healthcare, Uppsala, Sweden). After immunoprecipitation, the beads were recovered and boiled in SDS-containing gel sample buffer under reducing conditions to release the FV heavy chain. Samples were then subjected to gel electrophoresis and Western blotting. Bands were visualized using a monoclonal anti-FV heavy chain antibody (AHV-5146; Haematologic Technologies) and chemiluminescence. During this procedure, samples from the proband (but not his parents’ samples) were concentrated six times to increase the chances of detecting any residual FV(a).

Measurement of TFPI levels

Plasma levels of free TFPI were determined using a commercial ELISA kit (Asserachrom, Diagnostica Stago, Asnières sur Seine, France).

Thromboelastometry

Rotation thromboelastometry in whole blood was carried out on a ROTEM® Analyzer (Tem International GmbH, Munich, Germany) according to the manufacturer’s instructions. Coagulation was initiated with the extrinsic (EXTEM) or intrinsic (INTEM) trigger provided by the manufacturer. The following parameters were derived from the thromboelastogram: clotting time (CT), clot formation time (CFT), alpha-angle (α), maximum clot firmness (MCF), maximal velocity (MAXV) and area under the curve (AUC).

Genetic analysis

DNA mutation screening  Genomic DNA from the patient and his parents was isolated from peripheral blood leukocytes using the Wizard® Genomic DNA Purification kit (Promega, Madison, WI, USA). All 25 exons, the proximal promoter (approximately 1000 bp) and the 3′-UTR of the F5 gene were amplified by polymerase chain reaction (PCR) and sequenced using the BigDye® Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, Bedford, MA, USA), essentially as previously described [8].

Multiplex ligation-dependent probe amplification (MLPA) analysis  To check for possible large deletions (or duplications) within the F5 gene, MLPA analysis was performed. This technique [26] makes it possible to ‘count’ the copies of specific target sequences in a person’s genomic DNA by comparison with control sequences present in two copies in the same genomic DNA. In the absence of a commercial F5 MLPA kit, two sets of 11 F5-specific probes with lengths between 88 and 148 nt were designed and ordered from Integrated DNA Technologies (Leuven, Belgium). The first set (probemix A) included probes for exons 1, 4, 7, 10, 13 (proximal portion), 16, 19, 22 and 24, as well as intron 3 and the 3′-UTR of the F5 gene. The second set (probemix B) included probes for the promoter region as well as intron 2 and exons 6, 9, 11, 13 (distal portion), 14, 15, 17, 18 and 21 of the F5 gene. Each F5-specific probemix was combined with the control probemix from the SALSA MLPA kit P200-A1 Human DNA Reference 1 (MRC Holland, Amsterdam, the Netherlands), which contains 14 probes (172–250 nt in length) recognizing control genes on various chromosomes. Genomic DNA from the proband, his parents and four normal controls was standardized to 20 ng μL−1 and MLPA reactions were carried out according to the manufacturer’s instructions. MLPA products were separated on an ABI 3730 DNA Analyzer (Applied Biosystems) and results were analysed using the Peak Scanner software.

cDNA analysis  Although FV is synthesized in the liver, F5 cDNA analysis was conducted on RNA extracted from platelets, which are readily accessible and known to contain (ectopic) F5 mRNA. Total platelet RNA from the patient, his parents and two normal controls was isolated using TRIzol® Reagent (Invitrogen, Breda, the Netherlands). Briefly, 20 mL ACD-anticoagulated blood were centrifuged at 100 × g for 10 min. The resulting PRP was collected and centrifuged again at 1000 × g for 10 min to precipitate platelets. After discarding the supernatant, the platelet pellet was resuspended in 2 mL TRIzol® Reagent. Further steps were taken according to the manufacturer’s instructions. The washed RNA pellet was eventually resuspended in RNase-free water and quantified spectrophotometrically. Total RNA was subsequently reverse-transcribed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Ten overlapping fragments spanning the whole F5 mRNA were amplified from the obtained cDNA and analysed by agarose gel electrophoresis and direct sequencing. Primers and conditions are available on request. Gel bands were evaluated by densitometric analysis using the UN-SCAN-IT gel Version 6.1 software (Silk Scientific, Orem, UT, USA).

Bioinformatic analysis

Splice site consensus values for the sequences flanking the intron 8 pseudo-exon were evaluated with three bio-informatic tools: nnsplice version 0.9 (http://www.fruitfly.org/seq_tools/splice.html), SpliceView (http://zeus2.itb.cnr.it/~webgene/wwwspliceview_ex.html) and Human Splicing Finder (http://www.umd.be/HSF/) [27]. The latter program was also used to identify and score potential branch-point sequences.

Results

Case history

The proband is a 31-year-old Caucasian male from north-eastern Italy. The diagnosis of severe FV deficiency (FV:C < 1%) was made at birth, when he presented with numerous hematomas of the limbs and prolonged bleeding after resection of the umbilical cord. At the age of 1 month, he developed a hand hematoma that required administration of fresh frozen plasma. Throughout childhood he experienced recurrent epistaxis and ecchymoses of the limbs, which were treated with antifibrinolytic agents (tranexamic acid). From the age of approximately 10 years, he has been suffering from hemarthrosis (two to three episodes year−1, especially at the knees, shoulders and right elbow) and muscle hematomas following even minor traumas. The patient has undergone two minor surgical operations (excision of a cyst from the oral cavity and removal of a birthmark from the shoulder), which were successfully managed with prophylactic administration of fresh frozen plasma. However, he is not on routine prophylaxis; he is only treated on demand. No plasma or platelets were administered to him in the 2 months preceding each blood sampling for this study.

Recently, the patient developed a syncope associated with hemorrhagic shock (hemoglobin level of 4.5 g dL−1; normal range, 14.6–17.7 g dL−1) and he was admitted to the Intensive Care Unit, where a diagnosis of spontaneous right pneumothorax with hemothorax was made. Following the administration of large amounts of plasma and red blood cells and the insertion of a drainage tube into the right pleural cavity, he progressively recovered from the pneumothorax. Plasma was administered for several days, without bleeding recurrences, and the patient was eventually discharged 1 month after admission in good clinical conditions. However, 15 days later he reported to the Emergency Room because of a sudden headache and blurred vision in the left eye. A brain computed tomography scan revealed a subdural frontal intracranial hemorrhage, which was treated conservatively with the administration of plasma on admission and during the following days. Partial reabsorption of the intracranial hemorrhage was observed at day 10, and almost complete resolution at day 30. The patient was discharged and has had no relapses up to now.

The proband’s parents are reportedly unrelated. Both have reduced FV levels (FV:C 42% in the mother and 79% in the father), but no history of bleeding.

Thrombin generation

Thrombin generation was measured in PPP and PRP from the proband, his parents, one of the previously investigated FV-deficient patients (PD II) [24] and a normal control after triggering coagulation with 1, 5, 10 or 50 pm TF (Fig. 1). Thrombin generation in control PPP and PRP was already measurable at 1 pm TF and peak height increased (and lag time decreased) at higher TF concentrations. Moreover, platelet preactivation with collagen or Ca2+-ionophore greatly enhanced thrombin generation in PRP. Thrombin generation in the proband’s parents was similar or even higher (especially in the mother) than in control plasma. In contrast, no thrombin generation could be detected in the proband’s PPP or PRP at any of the TF concentrations used, not even after preactivation of platelets with collagen or Ca2+-ionophore. In this respect, the proband of the present study is different from patient PD II (Fig. 1) and the other two previously investigated FV-deficient patients with undetectable plasma FV (PD I and PD III), who all showed substantial thrombin generation in PRP [24]. The presence of an acquired antibody against FV in the proband’s plasma was excluded using a thrombin generation-based Bethesda assay (data not shown).

Figure 1.

 Thrombin generation in platelet-poor plasma (PPP) and platelet-rich plasma (PRP). Thrombin generation was measured in PPP and PRP from the proband, his parents, four normal controls and patient PD II [24] after activation of coagulation with 1 (light grey), 5 (middle grey), 10 (dark grey) or 50 (black) pm tissue factor (TF). Thrombin generation in PRP was determined in the absence of platelet agonists and after preactivation of platelets with collagen (10 μg mL−1) or Ca2+-ionophore (20 μm). Only one representative normal control is shown. In some panels, the thrombin generation curve evoked by 50 pm TF is not visible, because the lag time was too short to be measurable.

Genetic analysis

To identify the genetic defect(s) responsible for FV deficiency, the coding sequence (including splicing junctions), the proximal promoter and the 3′-UTR of the F5 gene were sequenced in the proband and his parents, but no mutation was found. Remarkably, however, the proband turned out to be homozygous for all known polymorphisms covered by the sequencing (n = 38), except one (327A/G in exon 2). This suggested that he might be hemizygous for a large portion of the F5 gene. To check for the presence of large deletions within the gene, MLPA analysis was performed using 22 F5-specific probes, but no abnormalities were noticed in the MLPA profiles (data not shown), strongly arguing against the presence of large F5 deletions in the family.

Alternatively, the proband might have inherited two identical alleles bearing the same mutation. As the coding region had already been screened without success, we reasoned that the causative mutation might be located in an intron and possibly affect splicing. To test this hypothesis, total platelet RNA from the proband, his parents and a normal control was isolated and reverse-transcribed to cDNA. When F5 cDNA fragments were amplified and analysed by agarose gel electrophoresis, a difference was noticed in the amplicon spanning exons 8–11 (Fig. 2A). While the normal control showed the expected 470-bp PCR product, the proband showed a higher-molecular-weight product, suggesting the retention of > 100 nt of intronic sequence in his mature F5 mRNA. Although hardly any normally spliced F5 cDNA was visible in the proband’s lane, densitometric analysis of the gel disclosed the presence of a very faint band comigrating with the control 470-bp band (Fig. 2B, arrow). The proband’s parents showed both the normal and the abnormal band. Interestingly, the abnormal band was less intense than the normal band, suggesting that the aberrant mRNA is subject to partial degradation in vivo. Sequencing of the abnormal amplification product revealed that the proband’s F5 cDNA contained a large (111 bp) insert between exons 8 and 9, whose sequence was identical to nucleotides 157–267 of intron 8 (Fig. 2C). Although the inserted sequence did not alter the mRNA reading frame, it contained an in-frame stop codon predicting premature termination of translation at codon 436 (within the A2 domain).

Figure 2.

F5 cDNA analysis. (A) Detection of the splicing abnormality. A F5 cDNA fragment spanning exons 8–11 was amplified in the proband (Pb), his parents (Mo, Fa) and a normal control (NC), and run on a 2% agarose gel. B, no cDNA control (blank). M, molecular weight marker. The 470-bp product represents the expected normal fragment, the 581-bp product visible in the proband and his parents is aberrant. (B) Densitometric scans of the proband’s, mother’s, father’s and normal control’s lanes of the gel shown in A. The small peak in the proband’s densitometric profile (arrow) indicates the presence of a very faint band comigrating with the control’s 470-bp band, which corresponds to the correctly spliced F5 cDNA. (C) Characterization of the aberrant mRNA (cDNA). The F5 pre-mRNA is spliced differently in the normal control (top) and in the proband (bottom). The sequencing chromatogram of the F5 8F-11R fragment amplified from the proband’s cDNA (primers are indicated by thick arrows) shows the insertion, between exons 8 and 9, of 111 bp derived from intron 8 (hatched). The insert contains an in-frame stop codon predicting premature termination of translation.

To understand why part of intron 8 is retained in the patient’s mature mRNA, three splice site prediction tools were employed to analyse the insert and the surrounding intronic sequence for the presence of splicing regulatory elements (Fig. 3). All three programmes identified a rather strong acceptor splice site consensus sequence (score between 0.83 and 0.90) at the 5′ end of the insert. Moreover, according to the Human Splicing Finder tool, which can also score branch-point sequences, a sequence (TGCTCAT, branch-point adenine underlined) with a branch-point consensus value of 0.90 was present 67–60 nt upstream of the 5′ end of the insert. In contrast, no donor splice site consensus sequence (two programmes) or only a weak one (score of 0.73, one programme) was identified at the 3′ end of the insert, which may explain why this intronic sequence is not normally included in the mature F5 mRNA. Following amplification (from genomic DNA) and sequencing of the relevant portion of intron 8, the proband was found to be homozygous for an A→G transition at nucleotide +268 of intron 8, whereas both parents were heterozygous (Fig. 3). Interestingly, the IVS8 +268A→G mutation affects the first nucleotide following the insert in the genomic sequence and creates a perfect donor splice site consensus sequence (score between 0.92 and 1.00) at the 3′ end of the insert, thereby causing exonization of this intronic region. These features make this portion of intron 8 a typical pseudo-exon [28] activated by the IVS8 +268A→G mutation.

Figure 3.

 Identification and bioinformatic analysis of the F5 IVS8 +268A→G mutation. Top left: amplification from genomic DNA of a 295-bp fragment spanning the pseudo-exon in F5 intron 8 (primers are indicated by thick arrows, gel lanes are labeled as in Fig. 2A). Top right: sequencing chromatograms showing the IVS8 +268A→G substitution in the proband (homozygous) and his parents (both heterozygous) vs. a normal control. Bottom: splice site and branch-point sequence consensus values for the intronic sequences flanking the pseudo-exon in intron 8. Splice site consensus values were determined with three different splice site prediction tools, whereas branch-point sequence consensus values could only be scored with Human Splicing Finder. The consensus values of the canonical splice sites at either end of intron 8 and of the canonical branch-point sequence at the 3′ end of intron 8 are also shown for comparison. Consensus values range from 0 to 1.00. UD, undetectable.

Polymorphism analysis indicated that the proband’s mother, who was heterozygous for the IVS8 +268A→G mutation, carried the R2 haplotype [29] on the ‘normal’F5 allele.

FV levels in plasma and platelets

To verify whether the F5 IVS8 +268A→G mutation is compatible with the expression of any (functional) FV, plasma and platelet FV antigen and activity levels were measured in the proband and his parents. Normal pooled plasma and a pool of platelets from 20 healthy individuals were used as references. FV-deficient patient PD II from our previous study [24] was also included as an additional control.

FV antigen, as determined by ELISA, was detectable both in plasma and platelets from patient PD II, but only in platelets from the proband. The proband’s parents had plasma and platelet FV antigen levels compatible with their heterozygous FV deficiency (Table 1).

Table 1.  Factor V antigen and activity levels in plasma and platelets
 Plasma FV antigen (%)Plasma FV activity (%)Platelet FV antigen (%)Platelet FV activity (%)
  1. UD, undetectable.

Proband< 2.0< 0.56.6UD
Mother43.037.233.826.5
Father67.965.143.031.5
PDII6.3< 0.51.73.3

FV activity levels were measured using a prothrombinase-based assay. No FV activity was detectable in the proband’s plasma or platelets, whereas the platelets of patient PD II expressed approximately 3.3% FV activity. The proband’s parents had reduced FV activity levels in both plasma and platelets (Table 1).

The FV heavy chain was also visualized on Western blot after immunoprecipitation from plasma and activated platelet suspensions (Fig. 4). No FV heavy chain was detectable in plasma from the proband or patient PD II (Fig. 4, left panel). In contrast, platelet FV heavy chain fragments were visible in both patients, although the signal was lower in the proband than in patient PD II (Fig. 4, right panel).

Figure 4.

 Western blot analysis of Factor (F) Va after immunoprecipitation. The heavy chain of FV was immunoprecipitated from plasma (left) and activated platelet suspensions (right), concentrated six times (proband’s and patient PD II’s samples) or not concentrated (parents’ samples), and subjected to gel electrophoresis and Western blotting. The control sample was run at different dilutions (100%, 50%, 25% and 5%). The FV heavy chain was detected using a monoclonal anti-FV heavy chain antibody and chemiluminescence. The molecular weight of the FV heavy chain is 105 kDa, the approximately 75-kDa band visible in platelet FVa from all individuals is a typical degradation product [24].

TFPI levels

Free TFPI levels in plasma were 22% in the proband, 43% in the mother and 46% in the father, in line with the respective FV levels [23].

Thromboelastometry

The overall coagulation phenotype of the proband and his parents was evaluated by rotation thromboelastometry in whole blood activated with either TF (EXTEM) or an intrinsic trigger (INTEM). Five healthy individuals served as controls. As shown in Table 2, the proband had markedly prolonged clotting time and clot formation time. Nonetheless a clot was formed and maximal clot firmness was similar to that of normal controls, both after extrinsic and intrinsic activation of coagulation. The proband’s parents had thromboelastogram parameters similar to those of normal controls. The mother had somewhat prolonged clotting time and clot formation time after intrinsic activation (probably attributable to her reduced FXII level: 33%), but once started, clot formation proceeded even more efficiently than in the normal controls.

Table 2.  Thromboelastogram parameters
 Controls
(n = 5)
ProbandMotherFather
  1. CT, clotting time; CFT, clot formation time; α, alpha-angle; MCF, maximum clot firmness; MAXV, maximal velocity; AUC, area under the curve. Parameters determined in the normal controls are expressed as mean ± standard deviation.

EXTEM
 CT (s)58.8 ± 12.92915846
 CFT (s)80.8 ± 25.92473665
 α (°)74.2 ± 5.0488378
 MCF (mm)57.6 ± 4.5527556
 MAXV (mm s−1)16.6 ± 3.6103520
 AUC (mm)5746 ± 454518574145562
INTEM
 CT (s)161.2 ± 11.8798385156
 CFT (s)69.2 ± 16.08514353
 α (°)76.2 ± 2.7737079
 MCF (mm)59.8 ± 3.6627762
 MAXV (mm s−1)18.8 ± 3.6152222
 AUC (mm)5956 ± 364612278796144

Discussion

Patients with severe FV deficiency express a highly heterogeneous clinical phenotype, but the determinants of bleeding tendency other than FV levels are poorly understood [14]. In a previous study, we have shown that three patients with undetectable plasma FV and mild bleeding symptoms had residual FV in their platelets, which was sufficient to support thrombin generation in PRP and probably protected them from major bleeding [24]. We therefore speculated that patients with equally undetectable plasma FV and severe bleeding symptoms might have less or no platelet FV. In the present study we tested this hypothesis by characterizing a FV-deficient patient with a severe bleeding tendency.

This patient was diagnosed with FV deficiency at birth because of umbilical bleeding, which is a rare (3%) and potentially life-threatening manifestation of the disease [15,16]. Unlike the previously investigated FV-deficient patients, who experienced only mucosal or post-traumatic bleeding [24], he suffers from recurrent spontaneous joint and muscle bleedings which, though not infrequent (20–25%) among FV-deficient patients [15,17], represent a more serious bleeding manifestation [16]. Moreover, he recently developed a spontaneous hemothorax and an intracranial hemorrhage. Overall, diagnosis at birth, the very young age (1 month) at first plasma transfusion, the nature and frequency of his bleeding symptoms and the frequent need for substitutive treatment all point towards a severe bleeding diathesis.

Sequencing of the F5 proximal promoter, coding region, splicing junctions and 3′-UTR yielded no mutation, but showed the proband to be homozygous at virtually all polymorphic positions. After excluding carriership of a large F5 deletion, this strongly suggested that the patient had inherited two alleles identical-by-descent and that his parents were (distantly) related. As a matter of fact, both parents turned out to carry the same splicing mutation in intron 8 (F5 IVS8 +268A→G), whereas the proband was homozygous. Because this mutation is deep-intronic, it was not detected during the initial genetic screening and could only be identified because of its impact on mRNA splicing. The F5 IVS8 +268A→G mutation activates a cryptic donor splice site in intron 8, causing a 111-nt long pseudo-exon to be retained in the mature F5 mRNA. As the inserted sequence contains an in-frame stop codon, no full-length FV can be synthesized from the aberrant mRNA, which also appears to be largely degraded by nonsense-mediated decay in vivo. However, a tiny fraction of the proband’s F5 pre-mRNA is spliced correctly (Fig. 2B), allowing for the possibility that the patient has traces of functional FV.

To our knowledge, only 11 splicing mutations have been reported in the F5 gene so far [22, 30–37]. All are located close to a splicing junction and disrupt an existing splice site consensus sequence, leading either to exon skipping or to the activation of a nearby cryptic splice site. In contrast, the IVS8 +268A→G mutation is deep-intronic and results in the inclusion of a whole new exon (a pseudo-exon) in the mature mRNA. Although only a few such mutations have been described to date in hemostasis-related genes, notably in the FVIII gene [38] and in the fibrinogen gene cluster [39,40], several examples have been reported in other genes and mutational activation of intronic pseudo-exons is emerging as a common genetic mechanism of disease [28].

While no FV antigen or activity could be demonstrated in the proband’s plasma, his platelets contained traces of FV antigen but no detectable FV activity. Accordingly, no thrombin generation was observed in his PPP or PRP on two different occasions. This is in striking contrast to the previously studied patients with undetectable plasma FV and relatively mild bleeding tendencies, who all showed thrombin generation in PRP already at 1 pm or 5 pm TF [24]. Because the proband of the present study had free TFPI levels that were equally low to or even lower than those of the previously studied FV-deficient patients [24], his undetectable thrombin generation in PRP is solely attributable to the virtual absence of functional FV in his plasma and platelets. This is likely to be a consequence of his more severe genetic defect (splicing mutation) as compared with the previously investigated patients, who all carried missense mutations. In fact, while the effects of missense mutations can be abolished by occasional translation mistakes, splicing mutations produce grossly abnormal mRNA molecules that are subject to nonsense-mediated decay. Remarkably, almost all FV-deficient patients with homozygous splicing mutations have experienced severe bleeding manifestations, like intracranial bleeding [22,31,33] or repeated hemarthrosis [32], suggesting that the type of genetic defect might determine the amount of residual platelet FV and thus clinical outcome in patients with severe FV deficiency.

Despite undetectable thrombin generation in the proband’s PRP, thromboelastometry showed that his blood could form clots of normal size and strength, although clot formation was markedly delayed. These findings are not conflicting, because the thrombin concentration (a few nm) needed for plasma to clot is too low to be detectable in the thrombin generation assay. Therefore, the patient may have traces of functional FV which, though below the detection limit, do afford minimal thrombin generation and clot formation in vivo. This once more illustrates the very low FV requirement for viability [21,22].

The proband’s parents had FV antigen and activity levels compatible with their heterozygous FV deficiency, and thrombin generation and thromboelastographic profiles similar to those of normal controls. The reduced plasma TFPI levels associated with their partial FV deficiency [23,24] may explain their comparatively high thrombin generation at low TF. As a carrier of a F5-null mutation (IVS8 +268A→G) and of the F5 R2 haplotype on different alleles, the proband’s mother was ‘R2 pseudo-homozygous’. Accordingly, she had consistently lower FV levels than the father and particularly high thrombin generation.

In conclusion, we have characterized a FV-deficient patient with a severe bleeding diathesis. In contrast to the previously studied patients with equally undetectable plasma FV but mild bleeding symptoms, this patient had undetectable platelet FV activity and showed no thrombin generation in PRP, due to homozygosity for a splicing mutation that virtually precludes FV synthesis. In combination with our previous observations [24], these findings suggest that thrombin generation in PRP may help predict bleeding risk in patients with undetectable plasma FV. However, this conclusion needs to be validated in a larger group of patients with severe FV deficiency.

Acknowledgements

The authors would like to thank the proband and his parents for their willingness to participate in the study and to donate blood several times. C. Bulato, S. Gavasso and M. Fadin are gratefully acknowledged for their excellent technical assistance. This study was supported by a VIDI grant (nr. 917-76-312, to E. Castoldi) from the Dutch Organisation for Scientific Research (NWO).

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.

Ancillary