The Molecular Genetics of von Willebrand Disease

Quantitative and/or qualitative deficiency of von Willebrand factor (vWF) is associated with the most common inherited bleeding disease von Willebrand disease (vWD). vWD is a complex disease with clinical and genetic heterogeneity. Incomplete penetrance and variable expression due to genetic and environmental factors contribute to its complexity. vWD also has a complex molecular pathogenesis. Some vWF gene mutations are associated with the affected vWF biosynthesis and multimerization, whereas others are associated with increased clearance and functional impairment. Moreover, in addition to a particular mutation, type O blood may result in the more severe phenotype. The present review aimed to provide a summary of the current literature on the molecular genetics of vWD. Conflict of interest:None declared.


Introduction
Von Willebrand disease (vWD) was first described by Erik von Willebrand in a large Aland Islands (Swedish-speaking region of Finland) family in 1924 and was reported to be the most common inherited bleeding disorder in 1926 [1,2]. Its prevalence is estimated to be 0.6%-1.3%; however, 1 in 10,000 patients have significant bleeding [3][4][5]. The characteristic clinical symptoms of vWD affect both males and females, and include mucosal bleeding (including epistaxis), menorrhagia, and prolonged bleeding following trauma or surgery. Severely affected patients may also bleed into soft tissues and joints [6].
vWD is associated with qualitative and quantitative deficiency of von Willebrand factor (vWF). Clinical diagnosis of vWD is based on von Willebrand antigen (vWF:Ag), factor (F) VIII clotting activity (FVIII:C), and von Willebrand ristocetin cofactor activity (vWF:RCo). According to The International Society of Thrombosis and Haemostasis (ISTH), vWD is classified as type 1, type 2, and type 3 ( Table 1). Type 1 and type 3 vWD are characterized by partial and complete deficiency of vWF, respectively. Functional deficiency of vWF is characteristic of type 2 vWD, which is further classified as 2A, 2B, and 2M based on defective interaction with platelets, and as 2N based on defective binding to the FVIII molecule. Although ISTH classification of 1994 indicated that vWD is a disease associated with vWF gene mutations, the 2006 revised ISTH classification states that, "additional genes might influence the biosynthesis and stability of plasma vWF" [7].

von Willebrand Factor
vWF is a multimeric plasma glycoprotein essential for primary hemostasis that mediates platelet plug formation via adhesion at the site of injury. In addition, vWF protects FVIII in plasma from proteolytic degradation by noncovalently binding to it and transports FVIII to the site of coagulation [8]. vWF is produced in endothelial cells and megakaryocytes as a pre-pro-polypeptide (pre-pro-vWF) that is 2813 amino acids long. Pre-pro-vWF is composed of a signal sequence that is 22 amino acids long, a pro-peptide sequence 741 amino acids long, and a vWF monomer sequence 2050 amino acids long ( Figure 1) [9][10][11].
During vWF biosynthesis the signal sequence is cleaved off by a signal peptidase after it is directed to the endoplasmic reticulum. Pro-vWF monomer undergoes dimerization at the CK domain via disulfide bonding between cysteine residues. In addition to dimerization, vWF is glycosylated in the endoplasmic reticulum. The dimers are then transported to the Golgi apparatus where they undergo amino terminal multimerization, sulfation, and further glycosylation, and change carbohydrates into high mannose oligosaccharides. Finally, propeptide is cleaved from the vWF multimers by furin (PACE [paired basic amino acid cleaving enzyme]) and remains non-covalently associated with the vWF multimers; 95% of vWF multimers are secreted constitutively and 5% of plasma vWF multimers are stored in the Weibel-Palade bodies (WPBs) of endothelial cells and are secreted in a regulated manner upon stimulation by thrombin, fibrin, and histidine [4,9,[11][12][13][14][15][16].  the vWF plasma level, which is also affected by a variety of other factors, including age, blood type, stress, thyroid hormone, pregnancy, single-nucleotide polymorphisms (SNPs) in the vWF gene-including the promoter region and other genetic loci [30][31][32][33]. Recent studies reported that macrophages are involved in the removal of vWF/ FVIII complex in the liver and spleen, and that D'-D3 domains are implicated in the clearance, although the precise mechanism of vWF clearance remains unknown [34][35]. Blood type is a major genetic determinant of the vWF plasma level [30]; it was reported that individuals with type O blood have 25% less vWF due to increased susceptibility to cleavage by ADAMTS-13 [36].

Type 3 vWD
Type 3 vWD is characterized by the virtual absence of plasma vWF and a consequent decrease in the FVIII level to 10%. The frequency of type 3 vWD is between 0.5 and 5.3/1,000,000 individuals [37,38]. Type 3 vWF gene mutations are recessive and such patients are homozygous or compound heterozygous for the vWF gene mutation that creates a null allele [39,40]. Nonsense mutations are the most common of the wide range of type 3 mutations scattered throughout the gene, most of which are in exon 28 ( Figure 2). The most common nonsense mutation is R1659X in exon 28 [41]. Nonsense-mediated decay of the allele-specific mRNA is thought to be the molecular mechanism of the nonsense mutations [42].
Missense mutations are the second most common mutations, some of which result in the replacement of cysteine residues that might cause multimerization and secretion defects [37,43,44]. Deletions resulting from recombination events include a single exon deletion, multiple exon deletion, and whole gene deletion. The most common deletion in the vWF mutation database is c.2435delC in exon 18. Deletion of exons 4 and 5 is reported to be a recurrent deletion in the UK, and is associated with a Research has shown that the pro-peptide sequence acts as a chaperone to direct vWF multimers to WPBs [17]. Glycosylation of vWF is an important post-translational modification that protects vWF from proteolytic destruction, affects plasma clearance, and maintains vWF's multimeric structure and interaction with platelets and collagen [18][19][20]. vWF multimers range in size from a dimer (~500 kDa) to ultra large multimers (>10 x 10 6 kDa). vWF multimers released from WPBs are ultra large (UL-vWF) and are the most effective for maintaining hemostasis at the site of injury. Pro-peptide's Cys-X-X-Cys sequence, similar to the active site of protein disulfide isomerases, is thought to be important for vWF multimer formation [20,21].
UL-vWF multimers are proteolytically cleaved by vWF cleaving protease (ADAMTS-13) to physiologically active plasma vWF multimer size within the A2 domain (Y1605-M1606) [22,23]. The A2 domain serves as a shear sensor and its unfolding is necessary for proteolytic cleavage to expose the cleavage site [24]. In addition to the importance of the cleavage site in the A2 domain for ADAMTS-13 proteolysis, the residues C1669 and C1670 that form a disulfide bond are important for A2 domain folding [25]. Furthermore, studies have shown that polymorphisms in the A1 and A2 domains affect the efficiency of ADAMTS-13 cleavage [26].
The half-life of vWF in circulation is 12-20 h. The vWF plasma level ranges between 50 IU/dL and 200 IU/dL in the general population [29]. Twin studies reported that genetic factors are responsible for 60% of the variation in  (HMW-vWF) multimers in both plasma and platelets. Type 2A vWD patients have a low vWF:RCo to vWF:Ag ratio (<0.6) [37]. Type 2A mutations include missense, deletion, insertion, and frameshift mutations; 73% of these mutations are located within exon 28 ( Figure 3) and 90% of these mutations are missense mutations that are either recessive or dominant [40]. D2 domain mutations are recessive and prevent multimer formation; patients are either homozygous or compound heterozygous with a null allele. Mutations in the D3, A2, A1, and CK domains are dominant mutations. D3 and CK mutations inhibit multimerization and dimerization, respectively [50,51]. A2 and A1 mutations result in an increase in susceptibility to ADAMTS-13 proteolysis, defective biosynthesis, or intracellular retention [52,53]. Type 2A vWD mutations that cause defective biosynthesis, intracellular storage, and intracellular retention (e.g. N528S) are group I mutations [54][55][56]; other mutations that increase ADAMTS-13 sensitivity are group II mutations (e.g. L1505R) [57]. Group I mutations result in a more severe phenotype than do group II mutations and patients respond better to DDAVP treatment [58].

Type 2B vWD
Spontaneous and increased binding of vWF to GpIbα receptors on platelets due to dominant gain-of-function A1 domain mutations and the absence of HMW-vWF multimers in plasma are characteristic of type 2B vWD. Due to spontaneous platelet binding HMW-vWF multimers in plasma are proteolyzed by ADAMTS-13 [7,59]. Patients have a low VWF:RCo to VWF:Ag ratio (<0.6) and an increase in ristocetin-induced platelet aggregation (RIPA) at low dose ristocetin [37,60]. Thrombocytopenia is also observed in some type 2B vWD patients under stress conditions, such as pregnancy or infection, and after DDAVP use. Patients may have giant platelets and also suffer from common haplotype and founder effect [45]. Although it is a rare complication, development of alloantibodies against vWF is observed in association with large deletions and whole gene deletion (Table 2) [39,46,47]. Alu (short stretch of DNA)-mediated recombination, impaired secretion, and multimerization are the causative mechanisms for some of the deletions. In addition, gene conversion events between the vWF pseudogene and the vWF gene that result in multiple substitutions and a stop codon in the vWF gene are reported to be a common pathogenic mechanism in type 3 vWD patients [48,49].

Type 2 vWD
There is a functional deficiency of vWF in patients with type 2 vWD, which is further classified as 2A, 2B, and 2M based on defective interaction with platelets, and as 2N based on defective binding to the FVIII molecule.

Type 2A vWD
Type 2A vWD is characterized by defective platelet binding due to the absence of high molecular weight vWF

Type 2M vWD
Type 2M vWD is characterized by a defect in vWFplatelet binding due to dysfunctional HMW-vWF caused by vWF gene mutations, despite a quantitatively normal vWF multimeric structure. Type 2M mutations are dominant loss-of-function mutations predominantly located within the platelet GP1b binding A1 domain [7,37]; 93% of these mutations are missense mutations and the remainder are small in-frame deletions [41]. Type 2M mutations are fully penetrant and 75% occur in exon 28 of the vWF gene ( Figure 5). A Canadian cohort study reported that a vWF:RCo to vWF:Ag ratio <0.4 in type 2M vWD patients was strongly associated with A1 domain mutations [69]. There are also a small number of mutations in the A3 domain (S1731T, W1745C, and S1738A) that affect collagen attachment and cause mild bleeding. Despite the fact that type 2M vWD patients respond poorly to DDAVP treatment, patients with A3 domain mutations respond well to DDAVP [70,71].

Type 2N vWD
Type 2N vWD is characterized by markedly reduced or lack of vWF affinity FVIII binding. Recessive mutations in the vWF-FVIII binding domain result in the lack of FVIII binding and a disproportionate decrease in the FVIII:C level to between 0.05 and 0.30 IU/mL. The type 2N vWD phenotype is observed in patients that are homo-spontaneous platelet agglutination [61][62][63][64][65]. Clinical manifestation of type 2B vWD varies and patients with thrombocytopenia have the most severe phenotype [66]. Furthermore, mouse models of type 2B mutations R2306Q and V1316M show that both mutations and ADAMTS-13 determine the phenotype [59].
There are >50 type 2B mutation submissions in the vWF mutation database. Type 2B mutations are highly penetrant and are detected only between codons 1266 and 1461-the region of exon 28 of the vWF gene encoding A1 domain; 96% of these mutations are missense mutations, most of which are observed at mutation hotspot arginine codons at positions 1306 (R1306W/Q/L), 1308 (R1308C/P), and 1341 (R1341Q/P/L) (Figure 4) [41]. A recent type 2B genotype-phenotype study reported that V1316M mutation is associated with the most severe bleeding score, as compared to other type 2B mutations [66,67]. Moreover, some type 2B mutations are associated with dynamic changes in the vWF level in association with the platelet count. For example, it was reported that a patient with R1306W mutation had normalized high molecular weight multimers (HMWM), but severe thrombocytopenia and a decrease in HMWM after correction of the platelet count [68]. In contrast to classical type 2B mutations, P1266Q/L (New York/Malmö) and R1308L mutations do not affect multimer size and do not cause thrombocytopenia [66].   [39,41] mutation is associated with a less severe phenotype and a FVIII level of 20 IU/mL, and such patients do respond to DDAVP [78].

Type 1 vWD
Type 1 vWD is characterized by partial quantitative deficiency of functionally normal vWF. The level of vWF is reduced to between 5 and 50 IU/dL, without significant abnormalities in multimer structure [5]. It is generally inherited as autosomal dominant; however, its clinical diagnosis is complicated due to incomplete penetrance and variable expression of the vWF gene [79]. In addition, compound heterozygosity for type 3 or type 2N mutations influence the severity of the disease. Recent studies performed in the European Union, the UK, and Canada have provided some data on the molecular pathology of type 1 vWD and established that there is a genotype-phenotype correlation [80][81][82]; the vWF gene was analyzed in ~300 type 1 vWD patients in the 3 studies, which demonstrated that both allelic and locus heterogeneity should be considered to play a role in the molecular pathogenesis of type 1 vWD.
Many candidate mutations, including promoter, splice site, nonsense, missense, and small insertions, as well as deletions have been identified; 80% of the mutations zygous for the same FVIII binding mutation, compound heterozygous for 2 different FVIII binding mutations, or compound heterozygous for a FVIII binding mutation and a vWF null allele [37,[72][73][74]. vWF binds to FVIII through its D' domain and part of the D3 domain between residues Ser764 and Arg1035 encoded by exons 18-23 in the vWF gene [75]; however, mutations beyond the FVIII binding regions (from exon 23 to 27) are also associated with decreased FVIII binding (e.g., Q1053H and C1060R) [76,77].
In addition to FVIII binding impairment, type 2N mutations might also cause secretion and multimerization defects, especially cysteine mutations (C788R/Y, Y795C, and C804F) [78]. Moreover, the 2 pro-peptide mutations R760C and R763G sterically inhibit FVIII binding by preventing furin cleavage of pro-peptide, resulting in the formation of UL-vWF multimers [37,74]. Type 2N mutations occur primarily between exon 18 and 20 [41]; 95% of the mutations are missense mutations ( Figure 6). R816W and R854Q are the most common type 2N mutations. Type 2N mutations are highly penetrant and the level of FVIII in type 2N vWD patients is associated with the specific mutation. For example, R816W mutation leads to a severe decrease in the FVIII level (<10 IU/mL) and patients do not respond to DDAVP, whereas R854Q

Genetic testing in vWD
Genetic testing of patients with inherited diseases has an important role in expanding our understanding of the molecular pathology of such diseases, and in decreasing disease-related morbidity and mortality. For some inherited complex disorders, including maturity onset diabetes of the young (MODY), genetic testing is important for differentiating disease subtypes and determining the optimal treatment method. Moreover, prenatal genetic diagnosis is extremely important for decreasing the frequency of inherited diseases as well as limiting the psychological and economic consequences for patients and their families.
vWD is a complex inherited bleeding disorder with clinical and genetic heterogeneity. Incomplete penetrance and variable expression are the major roadblocks to clinical diagnosis. Clinical diagnosis of vWD is based on phenotypic data; however, high variation in assays or lower detection limit, particularly vWF:RCo, and unavailability of certain tests like vWF:FVIIIB (vWF:FVIII bonding assay) or multimer analysis would also lead to misdiagnosis or inefficient diagnosis of vWD. Genetic testing of patients with vWD is based on vWF gene analysis. The value of genetic testing in vWD depends on the subtype; it is useful for the differential diagnosis and determining the proper treatment in patients with type 2 vWD. Genetic testing could be helpful in differentiating type 2N vWD from hemophilia A, which is possible by analyzing the exons encoding the FVIII binding region (exons [17][18][19][20][21][22][23][24][25]. Genetic testing could also be useful for differentiating type 2B vWD from platelet-type-vWD, which is based on analyzing just exon 28 in the vWF gene. In addition, genetic testing is important for the correct diagnosis of type 2A and type 2M vWD if multimer analysis cannot be performed. Genetic diagnosis of type 2A and type 2M vWD could also benefit the treatment of vWD, as type 2A patients respond to DDAVP, whereas type 2M patients do not. Clinical diagnosis of type 3 vWD is easily made based on phenotypic testing, as vWF is completely absent in the plasma. Nevertheless, genetic testing of type 3 vWD patients could be used for genetic counseling, prenatal diagnosis, and predicting inhibitor formation; however, the whole gene must be analyzed because mutations are scattered along the vWF gene. On the other hand, because correctly diagnosing type 1 vWD is clinically problematic and due to partial deficiency of vWF, molecular diagnosis is also problematic because of the complexity and mutational heterogeneity of the vWF gene. Many candidate mutations have been identified in type 1 vWD patients; in vitro expression studies are important for determining whether or not they are patho-are missense mutations and mutations primarily occur in exon 28 (Figure 7). In addition, some patients have more than vWF gene mutations. One of the major findings of these studies is that type 1 vWD is not always related to vWF gene mutations. Candidate mutations were identified only in 65% of patients and are more likely in patients with vWF:Ag <30 IU/dL. Moreover, mutation penetrance increases as the vWF plasma level decreases [37,81]. Type O blood type is associated with type 1 vWD in patients lacking any identified vWF gene mutation [39,81]. Studies that examined expression of the candidate mutations noted 2 primary pathogenetic mechanisms in type 1 vWD. The first mechanism is intracellular retention of the mutant vWF gene. Some vWF gene mutations, such as C1149R, were shown to dominantly impair vWF secretion [83,84]. Type 1 mutations that cause a loss or increase in cysteine residue (C2257S or G2441C) affect biosynthesis by causing significant intracellular retention and loss of multimeric structure. In contrast to this, D4 domain mutation L2207P caused similar significant intracellular retention and multimer loss to the mutations that involve Cys residue [85].
The second pathogenic mechanism is accelerated vWF clearance, which causes a very brief response to DDAVP in patients (<4 h), as compared to healthy individuals (6-9 h), and an increase in the vWF pro-peptide (vWFpp) to vWF:Ag ratio. Due to the clearance this type of vWD is also known as type 1C vWD. The mutations associated with accelerated vWF clearance are R1205H (vWD Vicenza), C1130G/F/R, W1144G, I1416N, and S1279F [5,[86][87][88].
Research has shown that ABO blood types also influence vWF clearance and the severity of the phenotype in vWF gene mutation carriers. For example, Y1584C mutation was the most common type 1 vWD mutation in 3 type 1 studies with incomplete penetrance. Although this mutation causes intracellular retention, all symptomatic Y1584C carriers also had type O blood in the Canadian and UK type 1 vWD studies, and patients in the UK study had an elevated vWFpp to vWF:Ag ratio [89][90][91][92]. Similarly, C2362F carriers with type O blood had a more severe phenotype [93].
There are also some common type 1 vWD mutations for which the molecular pathogenesis has yet to be discerned. For example, R924Q is a recurrent mutation associated with a founder haplotype and marks a splicing defect that created a null allele in a Canadian patient that was a compound heterozygous for R816W type 2N mutation; however, other studies reported that R924Q variation is a polymorphism [94,95]. genic variations. Hence, expression analysis of some candidate mutations showed they are just neutral polymorphisms. For some sequence variations, such as R924Q, the presence of a specific haplotype might be responsible for the disease phenotype. Moreover, it is likely that ≥35% of type 1 vWD patients do not have any vWF gene mutation. Finally, the presence of incomplete penetrance and the complex pathogenesis of vWD are major limitations to making a genotype-phenotype association in type 1 vWD patients. Consequently, although the use of genetic testing in type 1 vWD is of limited use, it could be used in patients with vWF:Ag <30% and in those with mutations that affect vWF clearance, such as R1205H mutation, for differentiating type 1 vWD from type 2 vWD [96][97][98].

Conclusion
vWD is an inherited bleeding disorder with a complex molecular pathology. Although numerous studies in various geographic regions have considerably advanced our understanding of the molecular mechanism of vWD, cases of vWD not associated with vWF gene defects are still observed. Complete understanding of the molecular pathogenesis of vWD requires additional in vitro expression studies that observe the effects of the candidate vWF gene mutations. In addition, use of whole genome or exome (part of genome formed by exons) sequencing (novel technologies) might identify other genetic determinants of vWD and help to complete our understanding of vWD by demonstrating the genotype-phenotype relationship.