Portal de Periódicos da CAPES

Clearance mechanisms of von Willebrand factor and factor VIII

2007; Elsevier BV; Volume: 5; Issue: 7 Linguagem: Inglês

10.1111/j.1538-7836.2007.02572.x

1538-7933

Peter J. Lenting, Carina J. M. van Schooten, Cécile V. Denis,

Hemophilia Treatment and Research

Coagulation factor (F) VIII and von Willebrand factor (VWF) are known for their crucial roles in hemostasis. The functional absence of either VWF or FVIII is associated with severe bleeding disorders, known respectively as von Willebrand disease (VWD) and hemophilia A [1, 2]. In contrast, epidemiological studies have revealed that elevated levels of VWF and/or FVIII predispose to thrombotic complications like venous thrombosis and stroke [3]. Both proteins circulate in plasma as a tight, non-covalently linked complex that is of major physiological relevance, as exemplified by the markedly reduced FVIII plasma levels observed in patients with no detectable VWF protein (VWD type 3) or with a defect in VWF–FVIII complex assembly (e.g. VWD type 2 N) [4-6]. Several mechanisms may explain this necessity for complex formation: (i) VWF stabilizes the heterodimeric structure of FVIII; (ii) VWF protects FVIII from proteolytic degradation by phospholipid-binding proteases like activated protein C and activated FX (FXa); (iii) VWF interferes with binding of FVIII to negatively charged phospholipid surfaces, which are for example exposed within activated platelets; (iv) VWF inhibits binding of FVIII to activated FIX (FIXa), thereby denying FVIII access to the FX-activating complex; and (v) VWF prevents the cellular uptake of FVIII. The intricate linkage between VWF and FVIII and its effect on the survival of these proteins has drawn considerable attention, not only because of its impact on hemophilia treatment via FVIII-replacement therapy, but also in view of the finding that abnormal clearance of VWF contributes to the pathogenesis of VWD. Being present as a complex complicates the assessment of how the VWF–FVIII complex or its individual components are removed from the circulation. It is only in the last few years that insight into the molecular basis of VWF–FVIII clearance has been obtained. The present review will address these recent findings. Pharmacokinetic studies using therapeutic FVIII preparations have revealed important information concerning the survival of FVIII in hemophiliacs and how the half-life of these preparations may vary between individuals [7-9]. However, it was only at the end of the 1990s that reports appeared providing the first insight into molecular mechanisms contributing to the removal of FVIII from the circulation (Fig. 1). Removal of factor (F) VIII, von Willebrand factor (VWF) and their complex in the liver. (A) Within the liver sinusoid, a number of different cell types can be distinguished: hepatocytes and stellate cells, which are separated from sinusoidal endothelial cells by the space of Disse. Macrophage-like Kupffer cells are incorporated between the sinusoidal endothelial cells. While circulating, VWF (orange) is in molar excess over FVIII (yellow). The majority of the FVIII molecules (approximately 95–98%) are found in complex with VWF, while not all VWF molecules are occupied with FVIII. (B) Free FVIII molecules are rapidly cleared via low-density lipoprotein (LDL) receptor-related protein (LRP1). Removal of free FVIII forces a constant shift from VWF-bound to free FVIII. LRP1-mediated clearance of unbound FVIII is facilitated by both LDL-receptor (LDLR) and heparan-sulfate proteoglycans (HSPG). These receptors also have the potential to internalize FVIII by themselves. Other potential receptors for FVIII are megalin, asialoglycoprotein receptor (ASGPR) and so far unidentified carbohydrate receptors. Each of these receptors is expressed on hepatocytes, while expression on macrophages has also been reported for some. VWF seems to be cleared via a pathway involving Kupffer cells, albeit that the responsible receptor(s) is/are currently unknown. When in complex with VWF, the uptake of FVIII by LRP1 is inhibited. Two groups simultaneously identified LDL receptor-related protein (LRP1, also known as α2-macrogobulin receptor or CD91) as a first candidate clearance receptor for FVIII [10, 11]. Interestingly, despite the efficiency by which LRP1 binds FVIII and mediates its transport to intracellular degradation pathways, the absence of LRP1 resulted in only a partial inhibition of FVIII degradation when addressed in cellular degradation assays [10, 11]. Apparently, LRP1-independent pathways must also contribute efficiently to the cellular uptake of FVIII. A body of evidence supports the in vivo relevance of LRP1 in the regulation of FVIII plasma levels. For example, a conditionally induced LRP1 deletion in MX1cre+LRP1flox/flox mice results in increased endogenous levels and prolonged survival of FVIII [12]. Epidemiological studies show that LRP1 also modulates FVIII plasma levels in humans [13-16], as evidenced by studies describing the relationship between two distinct LRP1 polymorphisms (LRP1/D2080 N and LRP1/A217 V) and increased FVIII plasma levels [14, 15]. Intriguingly, FVIII half-life is prolonged to a lesser extent when LRP1 is genetically deleted compared with when LRP1 is blocked via RAP [11, 12]. As RAP inhibits other members of the LDLR family, these members could also contribute to FVIII clearance. Bovenschen and colleagues elegantly showed that the archetype of this family, LDLR, interacts with FVIII and regulates FVIII plasma levels in a concerted fashion with LRP1 [17]. This cooperation is illustrated by the synergistic increase in FVIII survival upon deletion of both LDLR and LRP1 [17]. As for the other members of the LDLR family, no contribution of very-low-density lipoprotein receptor to the in vivo clearance of FVIII was found [18], and the contribution of other family members has not yet been documented. Heparan-sulfate proteoglycans seem to be a necessary element in the sequestration of FVIII at the cell surface in order to facilitate efficient binding to LRP1 [19]. Their in vivo relevance was demonstrated by coinjection of FVIII with the heparin-blocking agent protamine, resulting in a modest delay in FVIII clearance. Because heparan-sulfate proteoglycans can bind and internalize ligands themselves, albeit at a slower rate than LRP1 [20], they could also act as independent endocytic receptors for FVIII. Another recently discovered receptor that binds FVIII is the asialoglycoprotein receptor [21], an endocytic receptor recognizing terminal β-d-galactose or N-acetyl-d-galactosamine residues. In vitro experiments revealed virtually no binding of B-domainless FVIII to this receptor, suggesting that the FVIII B domain dominates interactions between receptor and FVIII [21]. As full-length FVIII and B-domainless FVIII have similar half-lives [22], the precise role of the asialoglycoprotein receptor in FVIII clearance in vivo remains to be assessed. Nevertheless, these data support a role of carbohydrate-recognizing receptors in the clearance of FVIII. Given its carbohydrate composition [23], members of the mannose receptor family, e.g. CD206, are candidate receptors. The molecular organization of FVIII comprises a heavy chain (A1-a1-A2-a2-B domains) that is non-covalently linked to the light chain (a3-A3-C1-C2 domains), and both chains contain LRP1 binding sites. One low-affinity interactive site is present within the FVIII C2 domain, while high-affinity interactive sites are located within the A2 and A3 domains, spanning residues Arg484–Phe509 and Glu1811–Lys1818, respectively [10, 11, 24, 25]. Of note, these regions not only overlap with interactive sites for FIXa and phospholipids, but are also well-known target epitopes for inhibitory anti-FVIII antibodies. This overlap might be a coincidence, as these regions are well-exposed at the molecular surface of FVIII and therefore fulfill criteria for interactive sites and inhibitor epitopes. Alternatively, recombinant LRP1 fragments are potent inhibitors of FVIIIa–FIXa-mediated FXa generation [26], pointing to a potential regulatory role of LRP1 in the down-regulation of the tenase complex. Are the respective LRP1 interactive sites accessible under physiological conditions? In their first report about FVIII and LRP1 interaction, Saenko and coworkers described the binding of FVIII A2 domain and intact heavy chain to LRP1 [11]. However, binding of the intact FVIII heavy chain was not observed in a study by one of us [10]. It was recently confirmed that intact heavy chain binds poorly, if at all, to LRP1 [27], explaining why site-directed mutagenesis of the LRP1 binding site in the A2 domain leaves in vivo pharmacokinetic parameters of mutated FVIII molecules unaffected [28]. In fact, the LRP1 interactive site within the A2 domain becomes exposed exclusively after proteolysis of the heavy chain [27]. This suggests that the LRP1–FVIII A2 domain interaction is part of the down-regulatory mechanisms that lead to inactivation at and/or removal of FVIIIa from the cellular surface. Exposure of LRP1 binding sites located in FVIII light chain is independent of proteolysis. Intact light chain, as well as its proteolyzed derivatives, displays similar binding to LRP1 [10, 27]. However, exposure of LRP1 interactive sites is also subject to regulation, albeit not via proteolysis. The main player in this regard is VWF, which interacts with both terminal ends of FVIII light chain [29]. This interaction is known to prevent binding of FVIII light chain to phospholipids and to FIXa. Given the overlap of interactive sites for FIXa, phospholipids and LRP1, it is not surprising that VWF interferes with the interaction between FVIII light chain and LRP1 [10]. VWF prevents binding of FVIII to LRP1 not only in a system using purified proteins but also in cellular degradation assays [10, 30]. Following proteolytic conversion of FVIII into FVIIIa, high-affinity binding to VWF is lost and LRP1 binding sites become exposed. Based on the above-mentioned mechanisms, it is tempting to conclude that FVIII survival is independent of LRP1 in the normal circulation: FVIII circulates in complex with VWF (thus protecting LRP1 binding sites in the light chain) with an intact heavy chain (thus with the LRP1 binding site in the A2 domain being encrypted). However, this conclusion opposes the in vivo observations in which deletion or blockage of LRP1 in mice with normal VWF leads to increased FVIII levels and FVIII survival. Thus, how does LRP1 contribute to FVIII clearance while FVIII is tightly bound to VWF? Possibly, the VWF–FVIII complex itself is cleared in vivo as a single entity via an LRP1-dependent mechanism. However, VWF does not interact with LRP1 [10], leaving only the option that LRP1 indirectly affects VWF clearance. Such a mechanism would explain why VWF levels are slightly, but significantly, increased in mice with a conditional LRP1 deletion [12]. Clearance studies monitoring the survival of VWF in these mice are needed to address this possibility. Sarafanov and coworkers have proposed an alternative pathway in which the VWF–FVIII complex is bound to heparan-sulfate proteoglycans at the cellular surface, leading to dissociation of the complex [19]. Once dissociated, FVIII is transferred to LRP1, whereas VWF is released into the circulation. The applicability of this pathway is probably limited to only a subset of cell types, because VWF completely blocks FVIII uptake in several cell types that have been tested in vitro [10, 30]. This potential limitation agrees with the modest effect of protamine on FVIII clearance [19]. A third aspect relates to the formation of the VWF–FVIII complex itself. The VWF–FVIII interaction is of high affinity (<1 nm) [31, 32]. Kinetic analysis revealed that complex formation represents a highly dynamic equilibrium, characterized by high association and dissociation rate constants (kon∼2 × 106 mol−1 s−1 and koff∼1 × 10−3 s−1) [31, 33]. As a consequence, a small but significant portion of the FVIII molecules (5–8%) circulate as a free protein [34, 35]. As such, FVIII light chain is unprotected by VWF, allowing LRP1-mediated clearance. Because free FVIII is cleared 4- to 6-fold more rapidly than VWF-bound FVIII [5], the loss of free FVIII is compensated by a continuous flow of VWF-bound to free FVIII in order to maintain equilibrium. But what happens with the remaining VWF-bound bulk (92–95%) of the FVIII molecules? Most probably, distinct pathways mediate clearance of FVIII vs. the VWF–FVIII complex. It seems conceivable that VWF-bound FVIII follows VWF in its clearance pathway. On average, the half-life of i.v. administered VWF antigen in VWD type 3 patients is about 15 h, but interindividual differences exist [36, 37]. Survival of endogenous VWF may also differ between individuals, as evidenced from studies using desmopressin treatment [38]. This agent induces the release of VWF stored in Weibel–Palade bodies, resulting in a temporary rise in VWF levels. The time it takes for VWF to return to baseline levels may be used to calculate the half-life of endogenous VWF. VWF propeptide is released from Weibel–Palade bodies simultaneously with mature VWF in a 1:1 molar ratio. However, survival of propeptide is 3- to 4-fold shorter compared with VWF, resulting in a distinct propeptide/VWF ratio under steady-state conditions [39]. A number of parameters can affect VWF half-life, although some of them may also influence VWF production. These parameters are discussed below. The average plasma level of VWF is ∼10 μg mL–1, although a broad range between individuals has been reported (from 40% to 240% of the population mean) [40]. One dominant factor influencing VWF plasma levels is its glycosylation profile [41]. Of importance, glycosylation also affects VWF function, an issue discussed in detail elsewhere [41, 42]. VWF contains 10 sites for O-linked and 12 sites for N-linked glycosylation. Part of the N-linked sugars contains ABO blood-group determinants [41], which are absent on the O-linked sugars as well as within the propeptide [43, 44]. Interestingly, the nature of the ABO determinant is strongly related to VWF levels: in persons with blood group O, average VWF concentrations are approximately 25% lower than in non-O individuals [42]. Levels are further reduced in persons with the Bombay phenotype, who lack expression of ABO antigens [45]. To explain the lower VWF levels in blood group O individuals, a number of possibilities have been proposed, such as impaired biosynthesis and secretion of VWF. This hypothesis has proven difficult to address, because glycan structures of VWF synthesized in vitro differ significantly from those of plasma VWF [46]. In vivo data do not favor an effect of ABO determinants on synthesis and/or secretion: total increase in VWF upon desmopressin treatment is similar for group O and non-O individuals [47]. A more probable explanation for reduced VWF levels in blood group O individuals involves a decreased survival. First, one would expect increased propeptide/VWF ratios in the blood group O population, as their VWF is cleared more rapidly but not their propeptide. Indeed, ratios were recently found to be consistently higher in the O group compared with the non-O group [43, 48]. Nossent et al. [49] used a mathematical approach and estimated that half-life of endogenous VWF is 2 h longer in the non-O population. Additional evidence that blood group O determinants on VWF are associated with increased clearance is provided by observations that infused FVIII (either plasma-derived or recombinant) disappears more rapidly in blood group O than in non-group O hemophilia A patients. Of course, this evidence seems only valid assuming that FVIII follows endogenous VWF in its clearance pathway. An interesting consequence of the blood group-dependent clearance rate is the possibility of developing plasma-derived VWF concentrates enriched in non-O-containing VWF molecules. Another direct link between VWF glycosylation and clearance is evident from studies using mice deficient in ST3Gal-IV. This enzyme mediates attachment of sialyl groups to terminal galactose residues. In ST3Gal-IV-deficient mice, the half-life of endogenous VWF is reduced 2-fold [50]. In addition, in a patient group referred to hospital for real or suspected bleeding disorders, reduced ST3Gal-IV-mediated sialylation was associated with reduced VWF plasma levels [50]. Thus, sialylation is important to prevent premature clearance via receptors that recognize non-sialylated terminal galactose residues, like the hepatic asialoglycoprotein receptor. Further support for the protective effect of sialyl groups comes from studies by Sodetz et al., [51] demonstrating that enzymatic removal of sialyl groups reduces the half-life of VWF in rabbits from 240 min to 5 min. The importance of the nature of terminal carbohydrate residues is most strikingly illustrated by RIIIS/J-mice. This strain is characterized by VWF levels that are severalfold lower compared with other strains, a phenotype because of a defect in a gene that is distinct from the VWF gene, i.e. Galgt2 (currently renamed as B4galnt2) [52, 53]. Murine expression of this glycosyltransferase is limited to intestine and kidney. In RIIIS/J mice, a genetic defect relieves this restriction and endows endothelial expression, allowing it to enrich VWF with a surplus of terminal N-acetylgalactosamine residues [53]. Upon secretion into plasma, this form of VWF is readily recognized by the asialoglycoprotein receptor and rapidly cleared. VWF is also subject to O-linked glycosylation. Stoddart and colleagues demonstrated that recombinant VWF lacking all O-linked carbohydrates has a reduced half-life when administered in rats [54], suggesting that O-linked sugars have a protective effect on VWF survival. Alternatively, suppression of O-linked glycosylation during synthesis may seriously compromise correct folding of the protein, therefore leading to the observed reduced half-life. The main O-linked glycan moiety (at least 70%) of VWF is composed of the sialylated tumor-associated T antigen [44]. In order to explore the potential connection between O-linked glycosylation and VWF levels, we have recently quantified the extent of glycosylation with sialylated T antigen. A negative relationship between the presence of the sialylated T antigen and VWF levels was observed in a large group of healthy individuals [55]. This association was most pronounced under pathological conditions, where VWF levels were outside the normal range, as illustrated by VWD type 1 patients where glycosylation with sialylated T-antigen was increased up to 3.6-fold compared with controls [55]. It remains to be investigated whether differences in O-linked glycosylation modulate the clearance rate. In support of this possibility is the observation that the propeptide/mature VWF ratio was increased 2-fold in the VWD type 1 patient group compared with normal individuals. Moreover, there was a strong linear correlation between this ratio and the extent of O-linked glycosylation [55], suggesting a decreased VWF survival in patients characterized by increased glycosylation with sialylated T antigen. When freshly released from storage granules, some of the VWF molecules can be classified as ultra- or unusually large (UL) multimers [56], a subclass of molecules that are able to interact spontaneously with the platelet glycoprotein Ib–IX–V complex [57]. To avoid the formation of VWF–platelet aggregates that may occlude the circulation, these UL multimers are subject to proteolysis by ADAMTS-13 (a disintegrin and metalloprotease with thrombospondin motif-13). The gene encoding this protease is located within chromosome region 9q34, remarkably close to the ABO locus (distance of about 140 000 bp) [58]. This chromosome region is an important genetic determinant of VWF levels, and it has been speculated that ADAMTS-13 cleavage leads to enhanced clearance of VWF [59]. However, current experimental data seem to oppose this hypothesis. Firstly, basal VWF levels are similar in normal and ADAMTS-13-deficient mice, when mice of the same genetic background are compared [60]. Levels should be increased in ADAMTS-13-deficient mice if proteolysis by this protease would enhance VWF clearance. Second, when wild-type (WT) VWF and a variant containing a type 2A mutation (making VWF more susceptible to ADAMTS-13 cleavage) were compared for their survival in a rat model, there was no difference in clearance rate [54]. ADAMTS-13 cleavage is therefore unlikely to participate in VWF clearance. VWD is classified into three major types: quantitative deficiencies are defined as type 1 (mild deficiency) or type 3 (severe deficiency), and qualitative effects are categorized as type 2 [61]. A large array of mutations in the VWF gene have been associated with VWD. Previous studies have revealed considerable insight into how these mutations may impair biosynthesis and/or secretion, thereby explaining low VWF levels in these patients. In recent years, it has become clear that increased clearance may also contribute to the pathogenesis of VWD. Casonato and coworkers provided the first report in this regard, describing the effect of desmopressin treatment in VWD patients with Vicenza subtype [62]. As expected, desmopressin treatment increased VWF to near-normal levels in these patients, but this freshly released VWF disappeared much more rapidly from their circulation compared with a control group, suggesting that the mutation, i.e. Arg1205 to His within the D3 domain, was associated with increased clearance [62]. To demonstrate the direct causative effect of this mutation, we tested the survival of the purified recombinant protein carrying the Arg1205His mutation in VWF-deficient mice [63]. The mean residence time (MRT) of this mutant was 0.3 h, compared with 2.8 h for WT VWF, clearly demonstrating that the survival of the mutated protein is severely reduced. As the description of the Arg1205His mutation as a cause for decreased survival of VWF, a number of other mutations have been described. These include Cys1130Phe and Cys1149Arg, located in the D3 domain and the CK domain residue Cys2671Tyr. In our experimental mouse model, these three recombinant mutant proteins displayed reduced survival (i.e. MRTs of approximately 0.7 h, which is about 4-fold shorter than that of WT VWF, but 2.3-fold longer than that of VWF/Arg1205His) [64]. Furthermore, in patients harboring these mutations, endogenous VWF disappeared 4.5-fold quicker compared with a control group following desmopressin treatment [64]. These patients were also characterized by an increased propeptide/VWF ratio, which is a surrogate marker for increased clearance. Monitoring the propeptide/VWF ratio is an approach that has been used in a more extensive manner by Haberichter and colleagues, who analyzed four families suspected as having VWD type 1 [48]. Within these families, all affected family members had a propeptide/VWF ratio outside the normal range and presented a strongly decreased survival of endogenous VWF after desmopressin treatment. Sequencing of their VWF gene revealed that two families had a Tyr1144 to Gly replacement in the D3 domain, and two families a Ser1179 to Phe mutation in the D4 domain. In total, six different mutations have now been identified that are associated with increased clearance, explaining at least in part the low VWF levels in these patients. This number suggests that increased clearance is a more common phenomenon than previously anticipated in the pathogenesis of VWD. The common occurrence was confirmed by Brown et al., [47] who measured VWF survival after desmopressin treatment in VWD type 1 patients compared with hemophilia A patients. The average survival of VWF in VWD type 1 patients was indeed significantly lower than in hemophilia A patients, although the report did not allow assessment of the individual data. How can so many different mutations be associated with increased clearance? No clear answer is yet available. Most of the mutations have been analyzed for their effect on VWF function, and no common defect was observed [63, 64]. This suggests that structural consequences of the mutations differ per mutation, and that there is no obvious structural denominator that makes the mutant proteins likely to exhibit decreased survival. To gain insight into the relative contribution of VWF regions to the clearance process, we have compared the in vivo survival of various recombinant VWF variants [63]. Our results show that (i) various regions within the VWF molecule contribute to the clearance process, (ii) the A1–A2–A3 region contains a receptor-binding site, and (iii) the regions D4–CK and D′–D3 comprise receptor-binding sites and/or regulatory sites that permit the enhanced or reduced exposure of the receptor-binding site within the A1–A2–A3 region. It is of further importance to note that four out of six mutations identified so far are located within the D′–D3 region of the VWF molecule. Alterations in this region may therefore increase the susceptibility of VWF to undergo accelerated clearance. In comparison with FVIII, little information is available regarding the identity of cells and receptors that mediate clearance of VWF. Various mechanisms can provide the basis for the removal of proteins from the circulation: interactions with one or more specific endocytic receptors; proteolysis; secretion via kidneys; and extravasation. For VWF, it is unlikely that (ADAMTS-13-mediated) proteolysis contributes to a significant extent. In the various experimental models, no proteolytic conversion of high VWF multimers into smaller derivatives has been observed [54, 63]. In view of its molecular size, it is also improbable that VWF is secreted via the kidneys or that it passes spontaneously from the circulation through the endothelium to the extravascular space. In contrast, its structural diversity allows VWF to interact with multiple receptors, thereby favoring the hypothesis of a receptor-mediated process. In search for the cellular basis of VWF clearance, we have monitored biodistribution of i.v. injected VWF and observed that the bulk of protein was targeted to the liver [63]. However, when expressed in relative terms, spleen was as efficient as liver in the uptake of VWF. Current studies are ongoing to identify the cells that are responsible for the removal of VWF and the VWF–FVIII complex. Preliminary data indicate that macrophage-like cells contribute to the clearance of VWF (P.J. Lenting, C.J.M. van Schooten, C.V. Denis, unpubl. data). So far, no endocytic receptors have been identified that mediate the removal of VWF from the circulation, with the exception of the asialoglycoprotein receptor. However, the reactivity of this receptor towards VWF is limited to cases of hyposialylation or altered glycosylation of VWF [50, 53]. Nevertheless, considering the strong influence of the glycosylation profile on VWF clearance, it seems conceivable that other carbohydrate-recognizing components play an important role in the removal of VWF from the circulation. Alternatively, we have recently reported that VWF may act as an adhesive surface for leukocytes through an interaction with β2 integrins [65]. In particular, αMβ2 integrin (also known as MAC-1 or CR3) is well known for its involvement in the uptake of microbes and proteins like fibrinogen by macrophages, and is therefore an attractive candidate to serve as an endocytic receptor for VWF. It is obvious that other candidates may exist, and it is the challenge for the near future to identify those that mediate the cellular uptake of VWF. In the past decade, important progress has been made in our understanding of the mechanisms that control circulating levels of FVIII and VWF. Although essential data are still lacking, such as the identity of VWF-clearance receptors, the current knowledge is already promising in that it may find application in the design of recombinant variants displaying prolonged survival. However, one should remain careful in the use of such therapeutic preparations. For instance, how will such modulations affect the cellular targeting of FVIII, VWF or its complex? It is not unthinkable that these modulations could result in a more efficient targeting to antigen-presenting cells, thus provoking an unwanted stronger immune response than that provoked by the currently available preparations. Apart from potential applications in the treatment of hemophilia, our understanding of the clearance mechanisms is of relevance for the diagnosis of certain VWD subtypes. Increased clearance of VWF is an aspect that is now recognized to be an important determinant of low VWF levels in VWD type 1. Haberichter et al. [66] have proposed that this phenotype be classified as VWD type IC (1-clearance), a proposal that is supported by us. Of course, increased clearance is not selective for VWF. We have for instance reported that low levels of protein S-Ser460Pro (Heerlen polymorphism) may be explained by a 4-fold reduced survival of this mutant. Whether or not FVIII mutations can be associated with increased clearance remains to be investigated. The authors state that they have no conflict of interest.