Molecular characterization, recombinant protein expression, and mRNA analysis of type 3 von Willebrand disease: Studies of an Italian cohort of 10 patients


  • Maria Solimando,

    1. Dipartimento di Medicina Interna, Università degli Studi di Milano, Milan, Italy
    Search for more papers by this author
  • Luciano Baronciani,

    Corresponding author
    1. U.O.S. Dipartimentale per la Diagnosi e la Terapia delle Coagulopatie, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano and Fondazione Luigi Villa, Milan, Italy
    • Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Via Pace 9, 20122 Milano, Italy
    Search for more papers by this author
  • Silvia La Marca,

    1. U.O.S. Dipartimentale per la Diagnosi e la Terapia delle Coagulopatie, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano and Fondazione Luigi Villa, Milan, Italy
    Search for more papers by this author
  • Giovanna Cozzi,

    1. U.O.S. Dipartimentale per la Diagnosi e la Terapia delle Coagulopatie, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano and Fondazione Luigi Villa, Milan, Italy
    Search for more papers by this author
  • Rosanna Asselta,

    1. Dipartimento di Biologia e Genetica per le Scienze Mediche, Università degli Studi di Milano, Milan, Italy
    Search for more papers by this author
  • Maria Teresa Canciani,

    1. U.O.S. Dipartimentale per la Diagnosi e la Terapia delle Coagulopatie, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano and Fondazione Luigi Villa, Milan, Italy
    Search for more papers by this author
  • Augusto B. Federici,

    1. Dipartimento di Medicina Interna, Università degli Studi di Milano, Milan, Italy
    Search for more papers by this author
  • Flora Peyvandi

    1. U.O.S. Dipartimentale per la Diagnosi e la Terapia delle Coagulopatie, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano and Fondazione Luigi Villa, Milan, Italy
    Search for more papers by this author

  • Conflict of interest: Nothing to report


Type 3 von Willebrand disease (VWD3) is characterized by unmeasurable von Willebrand factor (VWF) levels in plasma and platelets and severe but variable hemorrhagic symptoms. To identify and characterize the causal mutations, we screened 10 Italian patients with VWD3 by several techniques including Multiplex Ligation-dependent Probe Amplification to identify large insertions and deletions, High Resolution Melting and PCR coupled with Sanger sequencing. Fourteen different mutations scattered throughout the VWF gene were identified, 10 of which were novel. As expected, most of these mutations caused null alleles: five were deletions (del exons 1–3, del exon 17, c.2157delA, c.2269delCT, and c.3940delG), three nonsense (p.Q1526X, p.E1549X, and p.C2448X) and three potential splice-site mutations (c.658-2A>G, c.7729+7C>T, and c.8155+1G>T). Three candidate missense mutations (p.C2184S, p.C2212R, and p.C2325S) were also identified. Missense mutations and the putative splice-site defects were confirmed to be disease related by in vitro expression studies and mRNA analysis. None of these patients have developed alloantibodies against VWF. This study extends our previous finding that most of the mutations that we identified in VWD3 patients arise independently and are scattered throughout the entire VWF gene. Am. J. Hematol. 2012. © 2012 Wiley Periodicals, Inc.


von Willebrand factor (VWF) is a complex plasma glycoprotein, synthesized in endothelial cells and megakaryocytes, with a major role in primary hemostasis. It is required for platelet adhesion to damaged vascular endothelium, to bind, and stabilize coagulation factor VIII (FVIII) [1]. von Willebrand disease (VWD) is a frequent and clinically heterogeneous bleeding disorder due to quantitative (type 1 and type 3) or qualitative (type 2) defects of VWF. Type 3 VWD (VWD3), the most severe and rare form of VWD with prevalence of approximately 0.5–1 cases per million in the general population [2], is characterized by very low or undetectable levels of VWF in both platelets and plasma. VWD3 is associated with severe, but heterogeneous hemorrhagic clinical manifestations [3, 4]. Typically, patients are homozygous or compound heterozygous for null alleles due to nonsense mutations, small insertions/deletions, splice-site defects or, more rarely, large gene deletions spread throughout the VWF gene. Nevertheless, almost one-third of the identified VWD3 gene mutations are missense mutations (see; accessed January 2012). VWD3 patients may develop alloantibodies against VWF as a consequence of their treatment with VWF/FVIII concentrates. Since this complication renders replacement therapy ineffective and may also be the cause of life threatening anaphylactic reactions, screening for homozygous large deletions, associated previously with the presence of alloantibodies [5–8] is highly recommended. Of late, we have reported a retrospective studies on the management of inherited VWD in Italy that included 66 VWD3 patients. However, the patients reported in the current study have been collected subsequently to the above 66 patients and are unrelated to those cases. Following the data reported in the Registry on Hemophilia and Allied Disorders (update December 2010) [9], the known VWD3 Italian patients were up to 122 and more information on clinical and molecular markers on this population will be available in the future.

The aims of this study were to search for disease-causing VWF mutations in 10 VWD3 Italian patients using a combination of different techniques: Multiplex Ligation-dependent Probe Amplification (MLPA), High Resolution Melting (HRM) screening analysis, and PCR coupled with Sanger sequencing. The causative role of the identified missense mutations and splice-site defects were demonstrated respectively by in vitro expression studies and by mRNA analysis.



Patients were aware of the experimental nature of this study and gave informed consent, according to the Declaration of Helsinki (1964). Approval was obtained from Institutional Review Board of the Hemophilia Center of Milan, Bari, Florence, Perugia,and Reggio Calabria where the patients were recruited and followed-up. The bleeding history of all patients was collected according to a previously published bleeding score (BS) reported by Tosetto et al. [10].

Phenotypic study

Blood was drawn in 3.2% sodium citrate (9:1, v/v) and 5 mM EDTA (9:1, v/v). PAXgene™ Blood RNA tubes (PreAnalytiX, Hombrechtikon, Switzerland) were used to isolate total RNA from patients whole blood. FVIII coagulant (FVIII:C) and VWF antigen (VWF:Ag) levels were evaluated as described earlier [11]. Patients were diagnosed with VWD3 when plasma VWF:Ag was lower than 1 IU/dL. Alloantibodies to VWF were screened using previously described ELISA-based methods [12–14].

Mutation nomenclature and annotation

Nucleotides are numbered from the first adenine (+1) in the ATG initiation codon and amino acids are numbered from 1 to 2813, starting from the first methionine [15]. NM_000552.3 was used as a reference for nucleotide change whereas NP_000543.2 was used as protein reference sequence.

Methods used for the identification of VWF mutations

Genomic DNA was extracted from peripheral blood cells using standard methods [16]. All used oligonucleotides and technique conditions of the following methods are available on request. All patients were first tested for the presence of possible large deletions or duplications by MLPA. SALSA® MLPA® kits P011-B1 and P012-B1 VWF (MRC-Holland, Amsterdam, Netherlands) were used, following the manufacturer instructions, to investigate the 52 exons of VWF gene [17]. For each reaction, we used 150 ng of patient genomic DNA. Amplification products were separated by capillary electrophoresis using an ABI PRISM® 3130 Genetic Analyzer (Applied Biosystems, Foster City, CA) with the GeneScan™-500 LIZ® Size Standard (Applied Biosystems). Individual peaks, corresponding to each exon, were identified on the basis of the size standards by the GeneMapper® Software Version 4.0 (Applied Biosystems) whereas the relative signal of each probe was determined by the Coffalyser Software (MRC-Holland).

HRM analysis was then carried out for 25 exons (number 4, 6, 8, 9, 11, 12, 16, 17, 19, 20, 21, 23, 25, 30, 34, 35, 36, 38, 39, 40, 41, 44, 46, 48, 51) in all patients. This is a mutation scanning technique that monitors the progressive change in fluorescence caused by the release of an intercalating DNA dye from a DNA duplex as it is denatured with marginal increases in temperature [18]. HRM analysis has been limited to 25 exon for the presence of more then three polymorphisms in some of the exons or because of the presence of GC rich sequences. All 25 exons were amplified by PCR using the LightCycler® 480 High Resolution Melting Master kit (Roche Diagnostics, Mannheim, Germany). PCR reactions were performed in a 20-μL final volume containing 50 ng of genomic DNA following the producer guidelines. Samples were processed using a LightCycler® 480 Instrument (RocheDiagnostics). PCR and Sanger sequencing analysis was performed for all exons [8], starting with those where anomalous melting curves were observed during the HRM analysis, using an BigDye® Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems) after purification of PCR products with a QIAquick® PCR Purification Kit (Qiagen, Hilden, Germany). Purification by MultiScreen®HTS (Millipore, Billerica, MA) and Sephadex® G-50 (Sigma-Aldrich, Steinheim, Germany) was performed before loading the sample on an ABI PRISM® 3130 Genetic Analyzer (Applied Biosystems). The electropherogram sequences were manually read by more than one operator.

Breakpoint analysis

The exon 17 deletion breakpoints were determined by specific PCR assays. Using the online tool nucleotide BLAST ( we identified two homologous regions in introns 16 and 17. A PCR fragment including the hypothesized breakpoint region was obtained designing two primers, one matching at the 5′ of the homologous region in intron 16 (5′-GTGCTAGTGGTGGTCGTGAA-3′) and the other at the 3′ of homologous region in intron 17 (5′-GCACGGAAGGTAGGTCTGAG-3′). Sanger sequencing was subsequently performed on the successfully amplified region, to precisely define the deletion breakpoints. The PCR-amplified fragment in the presence of the deletion was 905bp, whereas in the absence of the deletion the expected fragment of 3699 bp was not obtained. The breakpoint of deletion of exon 1–3 was confirmed using the oligonucleotides reported by Mohl et al. [19].

RNA extraction and cDNA analysis

Total RNA from whole blood was isolated using the Paxgene™ Blood RNA System (PreAnalytix) following the manufacturer instructions in those patients carrying the potential splice-site defects (no. 2, 3, and 6) and in patient no. 7, where the second expected mutation was not found at the genomic DNA level. First-strand cDNA synthesis was performed starting from the total RNA, using random primers and the GoScript™ Reverse Transcription System KIT (Promega, Madison, WI) following the producer guidelines. DNA extending from cDNA nucleotides −40 to 1123 (exon 1–10) was amplified by PCR to verify the effect of the c.658-2A>G mutations (patient no. 3); the c.7729+7C>T mutation (patient no. 6) was evaluated by amplification of cDNA from nucleotides 7084 to 7853 (exon 42–47) and the c.8155+1G>T mutation (patient no. 2) was investigated by amplification of exons 47–52 (7787–8408). The obtained PCR bands were purified by QIAquick® PCR Purification Kit (Qiagen) and cloned into the pCR®-TOPO® Vector (TOPO® XL PCR Cloning Kit; Invitrogen, Carlsbad, CA) to be sequenced using the T7 primer. In patient no. 7, PCR and Sanger sequencing analysis was performed for all exons using cDNA oligonucleotides as reported above.

Plasmids construction

Vectors pSV-VWFH-C2184S (c.6651G>C), pSV-VWFH-C2212R (c.6634T>C), and pSV-VWFH-C2325S (c.6973T>A) were generated by site-directed mutagenesis (QuikChange® II XL Site-Directed Mutagenesis kit; Stratagene, Cedar Creek, TX), using as template the expression vector pSV-VWFH [11] and specifically designed oligonucleotides [20].

Transfection experiments

COS-7 cells were grown at full confluence into a six-well cell culture plate and were transfected with the appropriated amount of expression vector [wild-type (WT), mutant and both to generate mutant/WT recombinant VWF hybrids] using jetPEI™ transfection reagent (PolyPlus-transfection, Euroclone, Pero, Italy) as previously reported [20]. After 72 h of incubation, conditioned media and cell lysates were harvested and collected [11], recombinant VWF (rVWF) antigen levels were quantified and were expressed as mean values ± standard deviation. rVWF multimeric structure was analyzed by electrophoresis under nonreducing conditions in low-resolution gels as described before [11].


Phenotype and molecular results

All ten patients had clinical and laboratory features of typical VWD3, with high BS (median value 15, range 6–26) and undetectable plasma VWF levels (Table I). MLPA analysis identified two large deletions, one was found in the heterozygous state (del exons 1–3; c.1-30071_220+ 3445del) [19], whereas the other (del exon 17; c.2187-1920_2281+779del) was in the homozygous state. HRM analysis, allowed the identification of only two mutations (c.2157delA and p.C2325S) both in the heterozygous state. Sanger sequence analysis of the whole VWF coding region led to the identification of the remaining mutations. In total, 19 defective alleles were identified (all expected alterations but one) due to 14 different gene mutations, 10 novel and four previously reported (Table I).

Table I. Clinical, Biochemical, and Molecular Data of the Patients with VWD3
PatientSex/age M/F yearsVWF:Ag (IU/dL)FVIII:C (IU/dL)ConsanguineousBSNucleotide substitutionAmino acid substitutionExon (intron)Genotype
  • VWF:Ag, von Willebrand factor antigen; FVIII:C, factor VIII coagulant; BS, bleeding score; NR, not reported; NA, not available.

  • a

    Novel mutations.


Total RNA extraction and cDNA study

Reverse transcription PCR (RT-PCR) analysis of the platelet mRNA of patient no. 3 (c.658-2A>G), 6 (c.7729+ 7C>T), and 2 (c.8155+1G>T) identified three distinct splice-site alterations. These were: an homozygous deletion of 217 bp, due to the skipping of exon 7, in patient no. 3; an homozygous deletion of 40 bp, due to the skipping of exon 50, in patient no. 6; and an heterozygous insertion of five intronic nucleotides (GTAAG) between exons 45 and 46 in patient no. 2 due to the activation of a cryptic splice site in intron 45. RT-PCR analysis of the platelet mRNA of patient no. 7 did not show any splice-site alteration. All nucleotide substitutions identified in the heterozygous status in patients genomic DNA (mutation c.6551G>C and polymorphisms c.2555G/A, c.4141G/A, c.4414G/C, and c.4641C/T) were confirmed to be in the heterozygous status in the cDNA, sustaining the expression of the two alleles at the mRNA level.

Steady-state analysis of rVWF secretion

To evaluate the effects of the missense mutations on VWF synthesis and secretion, vectors pSV-VWFH-C2184S, pSV-VWFH-C2212R, and pSV-VWFH-C2325S were transiently transfected into COS-7 cells, either alone or in combination with pSV-VWFH, to generate mutant/WT rVWF hybrids. Three independent transfection experiments were performed in duplicate for the mutant, hybrid, and WT expression vectors. Secreted rVWF in conditioned media and intracellular rVWF in cell lysates were quantified as VWF:Ag, the amount of mutant and hybrids rVWF being expressed as a percentage of the WT rVWF taken as 100% (Fig. 1). C2184S (17 ± 3.4%), C2212R (21 ± 7.6%), and C2325S (35 ± 9.8%) rVWF in conditioned media were all remarkably reduced in comparison with the WT, whereas their corresponding hybrids showed intermediate values (C2184S/WT 55 ± 15.8%, C2212R/WT 89 ± 19.8%, and C2325S/WT 85 ± 39%). Increased values were found in cell lysates in both mutant (C2184S 251 ± 102.3%, C2212R 303 ± 67.8%, and C2325S 620 ± 90.8%) and hybrids rVWF (C2184S/WT 183 ± 63.1%, C2212R/WT 260 ± 84.1%, and C2325S/WT 153 ± 49.6%).

Figure 1.

Steady-state analysis of recombinant von Willebrand factor (rVWF) as expressed in COS-7 cells transfected with the appropriated plasmid DNA. rVWF values measured in cell media (upper panel), rVWF values measured in cell lysates (lower panel). Each bar represents average values of three independent transfection experiments performed in duplicate. The values are expressed as percentage relative to rVWF wild-type taken as 100%.

Multimer analysis of rVWF

The multimeric structure of mutant rVWF (C2184S, C2212R, and C2325S) and their hybrids in conditioned media was evaluated along with the WT. Secreted hybrids rVWF (C2184S/WT, C2212R/WT, and C2325S/WT) and WT showed a full range of multimers, whereas only dimers were present in all simple mutant VWF proteins (Fig. 2).

Figure 2.

Multimer analysis of wild-type (WT), mutant, and hybrid recombinant von Willebrand factor (rVWF) as expressed in COS-7 cells. rVWF WT, C2184S, C2212R, C2325S (panel A), C2184S/WT, C2212R/WT, and C2325S/WT (panel B) from conditioned media were electrophoresed in 1.2% agarose/0.1% sodium dodecylsulfate gel under nonreducing conditions. Lane NP shows normal plasma VWF. Multimers were visualized using a luminescent method [31].


Ten Italian patients with VWD3 have been analyzed and all of them, but one, have been fully characterized in this report. Five out of 10 patients were found to be homozygous, of these only three were known to be born from consanguineous marriages. A total of 14 distinct mutations were identified. Ten of them were novel and although none of these mutations was found in more than one patient, three out of the four previously reported mutations have been identified in VWD3 Italian patients [21]. Five were deletions: two large ones (del exons 1–3, del exon 17) and three resulting in the loss of few nucleotides. The deletion of exons 1–3 (c.1-30071_220+3445del) was reported to be the most common mutation in the VWD3 Hungarian patients [19]. Due to the fact that patient no. 9 is heterozygous for this deletion, and her parents were not available for the study, we could not establish the haplotype associated with this defect. However, all point polymorphisms identified in this patient were consistent with those reported by Mohl et al. [19] (data not shown). This finding, together with the patient Romanian extraction, strongly indicate a common origin with the deletion reported in the VWD3 Hungarian patients. The new deletion of exon 17 (c.2187-1920_2281+779del) was probably due to a recombination of two homologous sequences found in intron 16 and 17 (data not shown).

Only three nonsense mutations were identified (p.Q1526X, p.E1549X, and p.C2448X), and none was found to be involved in arginine codons (CGA>TGA). The low frequency of these mutations occurring in correspondence of the CpG dinucleotide in this group of patients is consistent with previous reports [21–27]. CpG dinucleotides are hotspot for mutation in the human genome [28]. However, only 11 CGA codons are present in the whole VWF cDNA sequence, whereas there are more than 1000 codons that can be changed to termination codons by a single nucleotide substitution. As previously reported, the nonsense mutations are likely to result into degradation of VWF mRNA due to nonsense mediated decay [29]. All three splice-site mutations (c.658-2A>G, c.7729+7C>T, and c.8155+1G>T) resulted in altering the normal mRNA splicing process leading to the following modified proteins respectively p.G220_S292delfsX27, p.G2706_ C2719delfsX25, and p.E2577GfsX21.

All identified missense mutations cause the loss of a cysteine residue p.C2184S, p.C2212R, and p.C2325S, the first two were located in the D4 domain, whereas the third one in the B1 domain. The expression study in COS-7 cells of the three candidate missense mutations confirmed their relationship with the VWD3 phenotype. Only small amounts of mutated rVWF (C2184S 17%, C2212R 21%, and C2325S 35%) were secreted by COS-7 cells in comparison to WT, whereas higher levels of mutated rVWF were present in cell lysates (C2184S 251%, C2212R 303%, and C2325S 620%). Only dimers were present in the conditioned media of cells transfected by mutant vectors, whereas co-expressed mutant/WT hybrid rVWF showed a full set of multimers but reduced secretions of 55% (C2184S/WT), 89% (C2212R/WT), and 85% (C2325S/WT) confirming the recessive role of these mutations. The inconsistency of the expression data with the patients phenotype might reflect differences between the in vitro expression system and the in vivo protein synthesis. The differences in the expression vector promoters and the type of cells used in our study might explain the high rate of protein productions induced in these in vitro experiments. A similar inconsistency has been reported for other missense mutations p.C275S, p.D141Y [11], and p.C2754W [30] found in VWD3 patients, all leading to the secretion of small amount of rVWF.

Two new techniques were used in this study to screen VWD3 patients rapidly and accurately for different types of mutations (large and small insertions/deletions, splice-site, and nonsynonymous-point mutations). MLPA analysis was very effective and allowed the identification of two different large deletions. One of these was in heterozygous state and therefore, it would not have been detected without this method. On the other hand, HRM analysis, which was used in this study for the first time in the molecular diagnosis of VWD3, was less efficient. Only two mutations were detected using this method out of the five identified by Sanger sequencing in the 25 exons screened. All three “missed” mutations, c.658-2A>G, c.2269delCT, and p.C2212R were in the homozygous state. HRM analysis is indeed less effective in detecting homozygous mutations: to overcome this problem, thus increasing the sensitivity of the technique, a possibility could be the addition of a wild-type genomic DNA to the patient sample. We have repeated the HRM analysis in the case of the three homozygous mutations using the patient genomic DNA jointly with a wild-type DNA, and two of these mutations (c.2269delCT and p.C2212R) were clearly detected (data not shown).

In one of our patient, the second expected mutation was not found, even though he was investigated by MLPA analysis and by Sanger sequence analysis of the whole coding region at both genomic and cDNA level. A possible explanation could be the presence of a deep-intronic mutation that causes an aberrant splicing not amplified by RT-PCR (e.g., introduction of a large pseudoexon in the mature transcript). The parents of patients with the mutations at the homozygous state, who were not aware of a possible consanguinity (no. 1 and 4), were investigated for the probands defects and confirmed to carry the expected mutation at the heterozygous state (data not shown).

Patients had severe hemorrhagic symptoms with the exception of those in childhood with milder BS. As expected, the cases showing the lowest levels of FVIII:C were indicatively those with the highest BS values. None of these patients were found to develop an inhibitor against VWF, not even patient no. 5 who is carrying the deletion of exon 17 in homozygous state.

In conclusion, the results of this study widen the mutational spectrum of VWD3 and confirm our previous finding that the majority of the mutations are responsible for null alleles and are scattered throughout the entire VWF gene. Indeed, none of these mutations was found in more than one patient, sustaining the allelic heterogeneity of the disease, although, among the four mutations previously identified, three were found in Italian VWD3 patients. Missense mutations, as previously reported, are more often than expected responsible of VWD3 and those identified in this study seem to act by causing the intracellular accumulation and degradation of mutated VWF.


The authors thank C. Ettore, A. Iorio, C. Latella, and M. Morfini for providing clinical data, plasma, and DNA samples of their patients. We acknowledge the illustration work of Luigi Flaminio Ghilardini. This work was supported by the Bayer Hemophilia Awards Program to Luciano Baronciani.

Authors Contributions: Maria Solimando carried out the molecular genetic studies, realization of expression vectors, and in vitro expression studies. Maria Solimando and Luciano Baronciani wrote the manuscript. Silvia La Marca, Giovanna Cozzi, and Maria Teresa Canciani were responsible for the phenotypic studies and characterization of recombinant proteins. Rosanna Asselta performed genetic screening analysis methods. Luciano Baronciani, Augusto B. Federici, and Flora Peyvandi conceived of the study, participated in its design and coordination, and helped to draft the manuscript. All authors read and approved the final manuscript.