Portal de Periódicos da CAPES

von Willebrand disease biology

2012; Wiley; Volume: 18; Issue: s4 Linguagem: Inglês

10.1111/j.1365-2516.2012.02840.x

1365-2516

Margareta Blombäck, Jeroen Eikenboom, David A. Lane, Cécile V. Denis, David Lillicrap,

Coagulation, Bradykinin, Polyphosphates, and Angioedema

von Willebrand's disease (VWD) is the most common inherited bleeding disease in humans. Since the first recognition of this disorder, considerable progress has been made in understanding the pathobiological mechanisms responsible for the enhanced bleeding exhibited by these patients. In this article, four aspects of VWD science will be summarized: a description of the original VWD discovery, a summary of current knowledge concerning the role of abnormal von Willebrand factor (VWF) storage and secretion in VWD, a biochemical characterization of VWF processing by ADAMTS13 and finally, a discussion of the role of mouse models of VWD in aiding our understanding of pathogenetic mechanisms. In 1926, Dr Erik von Willebrand of Helsinki reported in the literature several families who had, what he called 'hereditary pseudohemophilia'. Among these families were patients with Glanzmann thrombasthenia, some with essential thrombocytopenia, and some who had what we now call von Willebrand's disease (VWD) [1]. However, his article mainly focuses on the families he had investigated from the Åland Islands, especially the family 'S', who lived in the Norrgårds farm in Somboda, on Föglö Island (1-3). Map of the Åland Islands where Dr Erik von Willebrand first described the clinical features of von Willebrand disease and where the genetic transmission of the disorder was first characterized. Hereditary tree of family 'S' as described by EA von Willebrand in 1926 [1]. Photo of the family 'S' members as published in 'Recent investigations of the first bleeder family in Åland (Finland)' [9]. Augusta and Oscar 'S' were related and both had a history of nose bleeds, especially when they were younger; in both their families there had been 'bleeders'. Augusta had already given birth to 11 children when Dr von Willebrand first saw her in 1924. By then, three daughters had died of bleeding complications (two from gastrointestinal bleeds at 2 years old and one at age 4 after a tongue bite). Hjördis, at that time aged 5 years, had among other bleeding problems been in bed for 10 weeks after a laceration to her lip and had also experienced a bad ankle bleed. She had a much prolonged Duke's bleeding time (according to Duke), while her coagulation time and clot retraction were normal. Her capillary fragility test was positive and platelet numbers were normal. von Willebrand ended his paper by stating the finding of a positive capillary resistance test does not necessarily mean that there is an alteration in the capillary vessel walls. He thought the pathogenesis of the bleeding was caused by platelet dysfunction, in combination with a general lesion of the vessel wall. Hjördis bled to death at her fourth menstrual bleeding when she was 14 years old (3, 4). The gravestone of family 'S'. The index case of Dr von Willebrand's studies, Hjördis, died at the age of 14 from her fourth menstrual bleed. Four of her sisters died aged 2–5 years from bleeding. When we first met the family in 1957, Åland was a poor country recovering from the Second World War. Professor Inga Marie Nilsson, from Malmö, was a guest researcher at Professor Erik Jorpe's Medical Chemistry laboratory at Karolinska Institutet and I was employed at the laboratory. We had by then already investigated six Swedish families with pseudo-haemophilia and found that the probands and one parent, as well as several family members, had decreased levels of factor VIII, (FVIII) (at that time called AHG) and most had a prolonged bleeding time [2,3]. The probands had a very prolonged bleeding time and an AHG level between 1% and 10% of normal. Infusion of human fraction I-0 (a purified fraction of Cohn Fraction I) corrected both the bleeding time and AHG deficiency and the capillary bleedings. We deduced by different in vivo experiments that there must be a new factor [4–7]. I thought the probands were homozygotes for the trait, but it was not until Zimmerman and co-workers developed a rabbit antibody to the FVIII/VWF (FVIIIR:Ag) complex that real progress could be made. Many authors had by this time described similar patients with a severe haemorrhagic disorder (low AHG level and prolonged bleeding time) and had tried to find out if it was the same condition as that described by von Willebrand. Larrieu and Soulier suggested the name von Willebrand's syndrome, which later became VWD. We, however, considered it distinguishable from von Willebrand's thrombopathy or von Willebrand-Jurgens thrombopathy, as the German researchers preferred to call it [3]. However, Professor Jorpes came from Kökar (Fig. 1), another island in the Åland archipelago and he wanted us to investigate the Åland patients. So he took us to the city hospital of Mariehamn, Åland, where, in collaboration with Stig Arne Johansson and Birger Blombäck, we took blood samples from 15 members of family 'S' and several other families that von Willebrand had investigated. All severely affected patients had already died, and we found that those with mild bleeding symptoms had a decreased level of FVIII. We concluded that the Swedish patients with pseudo-haemophilia had the same disease as those on the Åland Islands [7]. In 1977, when reliable platelet function tests as well as an immunological assay of FVIII-related antigen (=VWF) were available, I returned with several co-workers to investigate four different Åland families thought to have VWD. In 1977, as in 1957, only patients with mild to very mild symptoms were available for investigation. It was found that the families could be divided into four categories [8]. Survivors with mild bleeding symptoms from family 'S' had the characteristics of VWD type 1 with similarly decreased levels of VWF antigen (FVIIIR:Ag) and ristocetin cofactor activity level, in addition to normal or decreased levels of FVIII. Their platelet aggregation was normal [8,9]. In one family, an isolated abnormal platelet aggregation was found when arachidonic acid was added, suggesting a special type of 'pure' cyclo-oxygenase defect. In a family from Lumparland, some had a mixture of the latter and VWD type 1 and others the platelet defect only. One family had a platelet dysfunction of the aspirin type i.e. the genuine cyclo-oxygenase defect [8–10]. During the 1980s, Dr Maria Anvret and I performed blood coagulation analysis in 25 type III (3) VWD patients from the haemophilia centre of Stockholm and DNA linkage analysis in nine probands and their families. Homozygosity and compound heterozygosity were suggested by the coagulation studies in the probands, and these results were confirmed by DNA linkage findings [11]. We also observed that the heterozygotes, which we called type 1, mostly had a bleeding tendency and an increased FVIII/VWFAg ratio (>1.6). Around 1990 Dr Zhiping Zhang from China started sequencing the whole VWF gene of the Swedish VWD type 3 patients of the Stockholm Center. He found that in two families with VWD originating from Finland, the probands were homozygous for a missense mutation in exon 28 and that 15 Swedish probands were homozygous for a single cytosine deletion in exon 18 [12,13]. Finally, in 1992, I made my third scientific trip to the Åland Islands, together with Zhiping Zhang and Dag Nyman. The exon 18 mutation was found in those with bleeding symptoms of family 'S', who were all heterozygous and in one homozygous boy from another related family. As shown in Fig. 5, no mutations were found in Gerd while her brother Lars, his son, and his grandson were heterozygous for the exon 18 deletion. As is shown in the figure, some of the family also had an exon 28 mutation. All five girls who died from uncontrolled bleeding were most probably homozygous for the exon 18 deletion. The heredity was in 1926 supposed by von Willebrand [1] to be dominant gender-linked. However, in 1957 [3] we suggested that the gene was autosomal dominant, which was confirmed by the pattern found by Zhang et al. in 1992 [13,14]. Mutation patterns of the families 'S' and 'I' members [12–14]. Half black: heterozygous for the deletion in exon 18. Half striped : heterozygous for the missense mutation in exon 28. Half black + half striped: compund heterozygous. Whole black: homozygous for the deletion in exon 18. Weibel-Palade bodies (WPB) are endothelial cell specific elongated secretory organelles that contain von Willebrand factor (VWF) and a variety of other proteins, including tissue-type plasminogen activator (tPA), P-selectin, interleukin-8 (IL-8) and angiopoietin-2. These mediators, which can be released from vascular endothelial cells upon stimulation of the cells by signalling molecules or mechanical stress, contribute to inflammation, angiogenesis and tissue repair (for an extensive review on WPBs see [15]). These organelles with a diameter of 0.1–0.3 μm and a length of 1–5 μm were first described in 1964 by Ewald Weibel and George Palade [16]. VWF is the major constituent of WPBs and is a prerequisite for the biogenesis of WPBs: endothelial cells of VWF-deficient animals lack WPB, whereas other non-endothelial cell types will form WPB-like organelles upon expression of recombinant VWF. During posttranslational modifications in the trans-Golgi network, VWF multimers are formed and are subsequently condensed into tubules that are targeted to WPBs [15,17]. Those tubules can be recognized by electron microscopy as the characteristic longitudinal striations in the WPB. Many secretagogues mediate release of WPBs, either by increasing intracellular free calcium (thrombin and histamine) or cAMP (epinephrine and vasopressin). Upon exocytosis, the VWF tubules unfurl into VWF strings that dock on the endothelial cells to mediate platelet adhesion. Three different modes of regulated exocytosis of WPBs have been described [15]: conventional exocytosis, in which single WPBs fuse with the plasma membrane and release their content; lingering-kiss exocytosis, where single WPBs fuse transiently with the plasma membrane via a small fusion pore and selectively release small molecules only but retain VWF [18]; and multigranular exocytosis, where several WPBs coalesce before exocytosis into large vesicles termed secretory pods [19]. When VWF is released into the blood it can form long strings and networks of strings that remain associated with the cells for some time and provide a platform for platelet adhesion. How the strings anchor to the plasma membrane is still a matter of debate, but integrin αvβ3 and P-selectin are potential candidates. Weibel-Palade bodies play a crucial role in the storage and timely secretion of VWF and defects in these processes may contribute to the phenotype of patients with von Willebrand's disease (VWD). The regulated secretion of VWF from WPBs can be stimulated with the synthetic vasopressin analogue 1-8 deamino-D-arginine vasopressin (DDAVP). DDAVP induces a prompt two to fourfold increase in VWF plasma concentration and is therefore an important treatment modality in patients with VWD. The effectiveness of DDAVP is very variable between patients and dependent on the type of genetic defect in VWF [20]. The effects of missense mutations in VWF on the formation and regulated secretion of WPBs are currently being studied and defects in the intracellular storage and regulated secretion of VWF seem to be a common mechanism underlying VWD. We have expressed recombinant wild-type and mutant VWF in a non-endothelial cell line, HEK293, which leads to the formation of so-called pseudo-WPB that resemble WPBs in endothelial cells [21]. Four missense mutations, located in the D3 and CK-domains of VWF and associated with a mainly quantitative deficiency of VWF, were expressed in HEK293 cells. All four mutations (p.Cys1060Tyr, p.Cys1149Arg, p.Cys2739Tyr and p.Cys2754Trp) diminished to some extent the storage in pseudo-WPBs, and led to retention of VWF within the endoplasmic reticulum. The pseudo-WPBs formed by mutant p.Cys1060Tyr are indistinguishable from wild-type VWF, as shown by immunofluorescence and electronmicroscopy data. The pseudo-WPBs formed by p.Cys1149Arg, p.Cys2739Tyr and p.Cys2754Trp are reduced in number, often short and sometimes round rather than cigar-shaped. However, other mutations in the D3 domain, causing type 2N VWD, have been reported to form normal rod-shaped storage organelles in HEK293 cells [15,22]. After incubation of the cells for 60 min with phorbol-12-myristate-13-acetate (PMA), which induces exocytosis of WPBs, the regulated secretion of VWF was shown to be impaired slightly for p.Cys1060Tyr but severely for p.Cys1149Arg, p.Cys2739Tyr, and p.Cys2754Trp. Co-transfection of wild-type and mutant VWF (to mimic the heterozygous state) partly restored the intracellular storage and regulated secretion of all mutants. From these data we conclude that defective intracellular storage and regulated secretion of VWF as a result of retention of VWF in the endoplasmic reticulum may be a common mechanism underlying VWD with a quantitative deficiency of VWF. As many of the missense mutations that reduce storage and secretion of VWF involve the loss of cysteine residues, we sought to determine whether the mutated cysteine's involvement in either an intrachain or interchain disulfide bond has a differential effect on the biogenesis of WPBs [23]. Three mutations were expressed in HEK293 cells: p.Cys1130Phe and p.Cys2671Tyr, which both disrupt intrachain disulfide bonds, and p.Cys2773Ser, which disrupts an interchain disulfide bond. The storage of VWF in pseudo-WPBs was reduced for the mutations p.Cys1130Phe and p.Cys2671Tyr and the mutant VWF was retained in the endoplasmic reticulum. Regulated secretion was also drastically impaired. However, the storage of the mutant p.Cys2773Ser was normal. Even though the mutation p.Cys2773Ser causes a severe dimerization and multimerization defect, resulting in mainly dimers and monomers, the mutant VWF was condensed into normal VWF tubules in the pseudo-WPBs. From these observations, we postulate that natural mutations of cysteines involved in the formation of interchain disulfide bonds do not affect the storage in WPBs and secretion of VWF, whereas mutations of cysteines forming intrachain disulfide bonds will lead to reduced VWF storage and secretion due to endoplasmic reticulum retention. Future experiments will focus on VWF string formation after WPB exocytosis and on the platelet adhesive properties of those VWF strings. Expression of VWF mutations in HEK293 cells is a valuable model to evaluate the pathogenic nature of VWF mutations at the cellular level. von Willebrand factor (VWF) is a large adhesive glycoprotein with established functions in haemostasis. It serves as a carrier for factor VIII and acts as a vascular damage sensor by attracting platelets to sites of vessel injury. VWF is a multidomain molecule that is assembled into multimers within the endothelial cell. It can be stored within Weibel-Palade bodies from where it can be released into the circulation. There is heterogeneity of molecular size of stored and released VWF. VWF size is important for its platelet adhesive function, with larger multimers being more haemostatically active. VWF in plasma may exist as multimers containing in excess of 100 monomer units. Functional imbalance in multimer size can affect phenotype: an increase in multimers can cause microvascular thrombosis, as in thrombotic thrombocytopenic purpura (TTP) whereas a reduction of very large multimers can lead to bleeding. Regulation of VWF multimeric size in plasma is carried out by the VWF-cleaving protease ADAMTS13 [24–26], a plasma metalloprotease that is constitutively active in the circulation. In recent years, much of the biology, biochemistry and pathophysiology of ADAMTS13 function has been clarified. In this section, we will focus on the biochemistry of VWF cleavage, a topic recently reviewed [27]. ADAMTS13 is a multidomain protease with metalloprotease, disintegrin-like, thrombospondin type 1 (TSP) repeats, cysteine-rich, spacer and CUB domains. ADAMTS13 activity is cation-dependent, with a reprolysin-like Zn2+ ion-binding signature (HEXXHXXGXXHD, single residue notation) involving three conserved His residues and an active site Glu225. Protease activity also requires Ca2+ ions that occupy a binding site within the metalloprotease domain and adjacent to the active site formed by Asp187, Asp182 and Glu212 [28]. Occupancy of the binding site appears to shape a loop that could potentially block the active site. Although several proteins are able to inhibit ADAMTS13 activity, there is as yet no evidence for physiological control of function by this means. Protease activity of ADAMTS13 in vivo is controlled therefore, not by natural plasma inhibitors, but rather by conformational changes in its substrate, which are induced when VWF is subject to elevated rheological shear forces [29]. Shear forces transform VWF from a globular to an elongated protein. This conformational transformation also unfolds the VWF A2 domain to reveal the scissile bond, Tyr1605-Met1606 (termed the P1-P1′ residues). VWF A2 domain unfolding is facilitated by an absence of an intradomain disulphide bond that connects the N- and C-terminal polypeptides of this domain, such as is present in the neighbouring A1 and A2 domains. These latter domains are consequently structurally resilient to shear force unfolding. Nevertheless, the VWF A2 domain does contain an important disulphide bond between adjacent Cys1669 and Cys1670 residues [30]. These residues form a very rare vicinal disulphide bond at the C-terminus of the last alpha helix of the VWF A2 domain. This vicinal disulphide bond forms a molecular plug interacting with hydrophobic residues in the core of the domain, which must be extracted for the domain to unfold [31]. Although VWF unfolding is essential for cleavage, not all binding interactions depend on unfolding. ADAMTS13 can bind to unfolded, globular VWF through an interaction mediated by the C-terminal TSP repeats and CUB domains of ADAMTS13 and the VWF D4-CK-domains [32,33]. This binding on its own does not result in proteolysis, but acts to position the protease should unfolding occur. Unfolding of the VWF A2 domain reveals cryptic exosites that interact with discreet sites on ADAMTS13, bringing the protease into position to enable it to perform its proteolytic function. These cryptic exosites include binding sites for the ADAMTS13 spacer and for the disintegrin-like domains. For the former, VWF residues Glu1660-Arg1668 at the C-terminus of the VWF A2 domain interact with ADAMTS13 spacer residues, of which Arg660, Tyr661 and Tyr665 are most important [34,35]. This interaction is of high affinity and contributes appreciably to the overall binding affinity between the two molecules. Interestingly, these surface-exposed ADAMTS13 spacer residues also serve as auto-antigens, generating auto-antibodies that cause the most common form of TTP, acquired TTP. For the latter, VWF Asp1614 (nine residues from the scissile bond) is proposed to interact with ADAMTS13 disintegrin-like domain residue Arg349 [36]. Although not of such high affinity, this interaction nevertheless helps the unfolding VWF A2 domain navigate the surface of the protease and contributes to cleavage efficiency. As the VWF scissile bond is brought into position over the ADAMTS13 active site, an essential interaction occurs with VWF residue, Leu1603, the P3 residue [37], which appears to make hydrophobic contact with a site (S3) on the ADAMTS13 protease domain. The S3 site has been proposed to comprise ADAMTS13 residues Leu198, Leu232 and Leu274. This helps locate the P1 and P1′ residues into their respective S1 and S1′ [38] binding pockets, bringing the scissile bond into position over ADAMTS13 Glu225 and allowing cleavage to proceed. von Willebrand's disease (VWD), caused by quantitative or qualitative abnormalities in VWF is considered the most common inherited bleeding disorder in humans. Mild and severe quantitative defects in VWF cause VWD type 1 and 3, respectively, whereas qualitative abnormalities induce VWD type 2. Murine models of VWD also exist whether engineered through gene targeting or as a result of naturally occurring mutations [39]. We will review briefly the various models reproducing the different subtypes of human VWD. The first VWD mouse model, the RIIIS/J strain, was identified because of a prolonged bleeding time caused by low VWF antigen levels. A common mutation in the VWF gene modifier B4galnt2, is responsible for the type 1 VWD phenotype in this mouse strain, as well as in a number of additional mouse strains. This mutation induces an increased clearance of the VWF protein, which is aberrantly glycosylated [40]. Alterations in other gene modifiers have been reported to lead to murine type 1 VWD. One such example relates to the deficiency in the ST3Gal-IV sialyltransferase, which leads to a dominant 50% reduction in VWF plasma levels and a prolonged tail bleeding time, also explained by increased clearance of the molecule [41]. More recently, hydrodynamic gene transfer has been used to generate mutation-specific type 1 VWD mouse models [42]. To this end, murine Vwf cDNAs carrying common type 1 VWD mutations identified in patients were injected into VWF-deficient mice via hydrodynamic injection. Interestingly, mice expressing the mutant VWF proteins reproduced the phenotype of the patients, validating such an approach to investigate the physiopathological mechanisms underlying type 1 VWD. No colonies of mice with type 2 VWD are currently available. The only models that have been reported are transient models for type 2B VWD generated via the hydrodynamic gene transfer approach [43,44]. Four different gain-of-function VWF mutantions identified in patients with type 2B VWD were expressed in the VWF-deficient mice, leading to a classical type 2B VWD phenotype: fluctuating thrombocytopenia, presence of platelet aggregates in the blood smears, abnormal multimeric pattern and defective haemostasis and thrombosis. Similar to human type 2B VWD, the severity of the phenotype was strongly mutation-dependent. Unfortunately, the limit of this approach where VWF is synthesized by transfected hepatocytes and secreted in the plasma did not allow a thorough investigation of other intriguing aspects of this VWD subtype such as abnormal megakaryopoiesis. A more stable model would be needed for this purpose. However, we have recently used a similar approach to generate transient murine models of type 2M VWD with abnormal collagen binding and again, a phenotype very similar to the patient's clinical data was obtained. The VWF-deficient mice generated through gene targeting represent a good model of type 3 VWD with no VWF detectable in any compartment, plasma, platelets, endothelial cells or subendothelium, factor VIII levels reduced by 80% and a strong haemorrhagic phenotype [45]. This model has been used extensively in research over the past 14 years, contributing not only to improve our knowledge about the role of VWF in thrombosis and haemostasis but also in other diseases where platelet adhesion and thrombus formation are involved [39]. These include but are not limited to, atherosclerosis, cancer metastasis, thrombotic thrombocytopenic purpura and stroke. A role for VWF in inflammation was also uncovered using this murine model, both directly through interaction with leukocytes and indirectly through the formation of Weibel-Palade bodies in endothelial cells and through regulation of the cell surface expression of P-selectin. Investigation of VWF clearance mechanisms and identification of VWF mutants leading to increased clearance was also made possible by the availability of the VWF-deficient mice [39]. von Willebrand's disease presents many interesting biological questions. Many details regarding the synthesis, storage and secretion and clearance of VWF, remain unresolved and although current therapies are safe and effective, improvements in clinical management are also needed. Overall, the biomedical and clinical interest stimulated by this condition will undoubtedly continue for sometime to come. The authors stated that they had no interests which might be perceived as posing a conflict or bias.