von Willebrand disease (VWD) is a genetically and clinically heterogeneous inherited hemorrhagic disorder caused by a deficiency or dysfunction of von Willebrand factor (VWF). VWD is considered to be the most common hereditary bleeding disorder [1, 2]. VWD is perhaps not uncommon in Taiwan, however, it has not been well studied. In type 2 and type 3 VWD, hemorrhagic symptoms are severer and laboratory data relatively more distinctive, which lead to diagnose the patient easily [3, 4]. Almost all the mutations detected in type 2 VWD including type 2A, 2B and 2M exhibit autosomal dominant inheritance patterns and is transmitted from either of their parents [3]. However, there was quite few de novo mutation of the type 2 VWD gene reported [5], somatic mosaicism of the type 2 VWD gene was even more rarely reported [3], despite it had been mentioned in hemophilia B [6]. Herein clinical, laboratory and genetic studies of only type 2 VWD exclusive of type 2N seen in Taiwan were investigated, which demonstrated a not-so-uncommon occurrence of de novo mutation. In this study we cited an interesting example from a big family suffered from type 2A VWD, in which somatic mosaicism of the VWF gene mutation will be presented.

Fifty-four patients from 23 unrelated families in whom type 2 VWD has been diagnosed not only by clinical and laboratory studies but also by genetic confirmation during 1990 to 2015 at National Taiwan University Hospital and Changhua Christian Hospital, all tertiary referred hospitals, were investigated. A revised classification of VWD described by Salder JE et al. in 2006 [7] and criteria for the diagnosis of type 2 VWD and its subtypes given by Federici AB in 2014 [8] were used. Factor VIII coagulant activity (VIII:C) was assayed according to the one-stage method. von Willebrand factor antigen (VWF:Ag) was determined by a two-step enzyme immunoassay sandwich method with a final fluorescent detection (ELFA) (VIDAS; bioMérieux, Marcy-I’Étoile, France) [9]. von Willebrand factor activity (VWF:Act) was measured by a latex particle enhanced immunoturbidimetric assay (Instrumentation Laboratory, Lexington, MA, USA). The latex particle was absorbed onto a specific anti-VWF monoclonal antibody directed against the platelet binding site of VWF, i.e. glycoprotein Ib receptor [10, 11]. VWF multimer analysis was performed as described previously [12]. A novel technique named amplification refractory mutation system-quantitative polymerase chain reaction (ARMS-qPCR), which was developed for both preimplantation genetic diagnosis and prenatal genetic diagnosis [13, 14], was used to confirm the presence of somatic mosaicism, because it can accurately quantify the number of copies of a particular amplicon, and determine the ratio of mutant to wild-type alleles in a sample, all using purified DNA as s test sample. Exons and its junction of full length VWF gene were amplified and sequenced as previously described [15]. Informed consent was obtained from patients and family members.

Seven out of 23 index patients (30.4 %) or 11out of 54 patients (20.4 %) available for examination showed novel mutations of VWF gene. With a large size of the VWF gene, this high degree of new variant can be expected. Although we consider these normal variants as novel mutations, their pathogenicity for VWD has to be confirmed by linkage analysis or structure-function studies of VWF gene, otherwise some nucleotide changes may be neutral in phenotype and can only be called as a polymorphism. In addition, the pathogenetic basis of some of the changes are also unclear.

De novo mutation has been identified in 4 families among 15 families (26.7 %) in whom family members including parents were available for examination. De novo mutation is a rare event in VWD [5]. However, some multicenter studies on type 1 VWD had identified de novo mutation in at least 2–4 % of proband [16–18]. Such high incidence (26.7 %) of de novo mutation in type 2 VWD patients exclusive of type 2N in the present study can be a selection bias or due to small number of patients studied. Two possible explanations for the occurrence of de novo mutation are a true event, or results of a new mutation coexisting with a germline mosaicism in a parent, or somatic mosaicism [18]. If more than one child is affected by so called de novo mutation with the same mutant, germline mosaicism can be rationally speculated, because the chance of de novo mutation with a same genetic defect in two children must be extremely rare [18, 19]. Only one child is affected in each family as found in the 4 families we identified (Table 1), the explanation for this may be a de novo mutation or a germline mutation with a transmissible mutant, or somatic mosaicism, unless we are able to obtain a genital tissue for confirmatory analysis. If there is a single affected offspring, germline mosaicism still cannot be excluded, this will depend upon the extent of the mosaic germ cell, a factor that is often impossible to quantify. Germline or gonadal mosaicism has been detected in type 2B VWD [19].

We have found an interesting example of somatic mosaicism of VWF gene in a large family with type 2A VWD with a previously reported VWF gene mutation. There are some key features of autosomal dominant inheritance, i.e., variable expressivity, incomplete or non-penetrance and anticipation. However, there is a general belief that type 2A, 2B and maybe 2M VWD mutations are fully penetrant [3], and are also not associated with variable expressivity. This is in marked contrast with type 1 VWD where all of these genetic variances are common for many mutations, and correlation between genotype and phenotype is difficult [3, 20]. Meanwhile, key features of an autosomal dominant inheritance pattern involves variable expression, incomplete penetrance and anticipation are not associated with type 2 forms of VWD [personal communication with Dr. David Lillicrap].

In this index family of VWD, the father carries the mutant allele but remains asymptomatic throughout his life, while all his children and grandchildren who carry the mutation have abnormal bleeding phenotype, namely patients of VWD. Somatic mosaicism can thus be attributed to this extraordinary presentation of this autosomal dominant disorder since the percentage of the VWF-producing progenitor cells of the vascular endothelium may be very low. From the patients and unaffected mother, the mutant allele versus the wild-type allele follows the classical 50 %:50 % and 0 %:100 %, respectively. However, either the real-time PCR data obtained from the asymptomatic “mutant carrier” were much deviated from 50 %, i.e., 25.5 % in the peripheral blood and 31.1 % in the oral mucosa. This is the undisputed evidence of the existence of “somatic mosaicism”. We therefore speculate that the cell/tissue which are responsible for coding VWF, e.g. vascular endothelial cell or megakaryocyte are carrying the wild-type alleles only, instead of mutant alleles. Such somatic mosaicism have been proved in other disorder such as Dravet syndrome [21 and hemophilia B [6].

Seven out of 23 index patients (30.4 %) examined showed novel mutations or normal variants of VWF gene. De novo mutation of type 2 VWD gene was identified in 4 families among 15 families (26.7 %) in whom family members were available for examination. Only one child was affected in each family. Germline mosaicism is therefore unlikely although its possibility cannot be totally excluded. Somatic mosaicism of VWF gene mutation was documented in a large family with type 2A VWD with a previously reported VWF gene pathogenic mutation, p.Arg 1597Trp. This extraordinary presentation was confirmed by a newly developed technique named ARMS qPCR.