Measurement of von Willebrand factor binding to a recombinant fragment of glycoprotein Ibα in an enzyme-linked immunosorbent assay-based method: performances in patients with type 2B von Willebrand disease
Claudine Caron, Laboratoire d'Hématologie, Hôpital Cardiologique, Bd du Professeur J. Leclercq, 59037 Lille Cedex, France. E-mail: firstname.lastname@example.org
Type 2B von Willebrand disease (VWD) is characterised by an increased affinity of von Willebrand factor (VWF) for its platelet receptor glycoprotein Ib (GPIb). This feature is usually studied in vitro by a ristocetin-dependent VWF platelet-binding assay, which has some limitations as it requires [e.g. (radio)-labelled anti-VWF antibodies and normal formaldehyde-fixed platelets]. We, here, extended the applicability of an enzyme-linked immunosorbent assay-based method previously described for the measurement of ristocetin co-factor activity that used a recombinant fragment of GPIb (rfGPIbα) and horseradish peroxidase-labelled rabbit anti-human VWF antibodies for measuring the captured ristocetin-VWF complexes on the rfGPIbα. Thirty-one type 2B VWD patients from 15 families with eight different known mutations were studied. VWF in plasma from 28 of these patients bound better than normal VWF at 0·2 mg/ml ristocetin, with the ratio, optical density (OD) patient/OD normal pool plasma, higher than 1·8. For two of the three other patients with no enhanced response of plasma VWF, the platelet lysate VWF showed an enhanced binding capacity; for the last patient, the results in other members of the family are unequivocal. We conclude that, this new method for measurement of plasma or platelet VWF-binding capacity offers great advantages for correct type 2B VWD diagnosis.
von Willebrand factor (VWF) is a large multimeric glycoprotein present in plasma, α-granules of platelets, Weibel–Palade bodies in endothelial cells and subendothelial matrices. It plays a major role in the process of platelet adhesion by interacting with both the subendothelium and the glycoprotein (GP) Ib receptor on platelets. In vivo, plasma VWF does not bind to platelet receptors, but its initial binding to collagen and/or other subendothelial matrix components in areas of high-blood shear stress, appears to be the trigger for the binding to GPIb. This VWF/GPIb interaction can be studied in vitro by using ristocetin as an agonist (Scott et al, 1991). The domain of VWF that interacts with the GPIb receptor is present on a trypsic fragment corresponding to amino acid residues V1212–K1491 (Fujimura et al, 1986). The cysteine loop (C1272–C1458) in the A1 domain of VWF, has been identified as an important element in the regulation of binding to GPIb. The region of GPIb involved in this interaction has been identified to the amino-terminal 42 kDa globular part of GPIbα (Vicente et al, 1990).
Defects in VWF result in von Willebrand disease (VWD), the most common inherited bleeding disease. The current classification (Sadler, 1994) distinguishes quantitative VWF defects, partial in VWD type 1 or complete in type 3, and qualitative ones in type 2, with four major subtypes (2A, 2B, 2M and 2N). Type 2N refers to variants with a markedly decreased affinity of VWF for factor VIII. Type 2A VWD refers to variants with decreased platelet-dependent function that is associated with the absence of high-molecular weight (HMW) VWF multimers, whereas type 2M refers to variants with decreased platelet-dependent function that is not caused by the absence of the HMW VWF multimers. In contrast, type 2B refers to variants with increased affinity of VWF for platelet GPIb generally resulting in spontaneous binding of VWF to platelets that form aggregates, detected as a loss of plasma HMW VWF multimers and mild thrombocytopenia. In vitro, this increased affinity allows a ristocetin-induced platelet aggregation (RIPA) in platelet-rich plasma (PRP) at lower concentrations of ristocetin than needed for normal PRP (Ruggeri et al, 1980).
The classification of a type 2 VWD (except subtype 2N) is first suggested by a reduced ristocetin co-factor activity of VWF (VWF:RCo) compared with the VWF antigen (VWF:Ag) level, with a ratio of VWF:RCo to VWF:Ag usually lower than 0·7, according to the guidelines recently proposed for Italy by Federici et al (2002). The measurement of VWF:Ag is easy to perform by well standardised and sensitive enzyme-linked immuosorbent assay (ELISA) methods using poly/monoclonal anti-VWF antibodies for which automated methods on a latex agglutination basis have also been developed, allowing rapid determination in urgent cases (Veyradier et al, 1999). The VWF:RCo classically is performed on normal platelets mixed with patient platelet poor plasma (PPP), but this is a time-consuming assay, based on the measurement of platelet agglutination in an aggregometer (MacFarlane et al, 1975) or by a rapid macroscopical slide technique (Brinkhous et al, 1975) using preferentially fixed platelets rather than fresh ones. Furthermore, this test is hampered by an unacceptable poor reproducibility. Recently, we described a more sensitive, reliable and reproducible ELISA method (Vanhoorelbeke et al, 2000) that uses a recombinant GPIbα fragment immobilised on a microtitre plate. To further classify a patient in a specific VWD type 2 subtype, more advanced investigations are required, such as the VWF collagen-binding (VWF:CB) assay, the plasma VWF multimeric analysis, and the RIPA assay. The RIPA assay enables the identification of patients that display an enhanced aggregation response at low ristocetin concentration. This paradoxal response is the main feature of type 2B VWD but also of platelet-type pseudo-VWD, where a gain-of-function of the GPIbα is present (Weiss et al, 1982). To differentiate type 2B VWD from pseudo-VWD, a direct measurement of the capacity of the patient plasma VWF to support ristocetin-induced aggregation of normal washed platelets, has been proposed (Weiss & Sussman, 1986). An indirect method measures, by ELISA, residual-free VWF present in the supernatants (Ruggeri et al, 1980). Furthermore, Meulien et al (1992) described a ristocetin-dependent platelet-binding assay in which recombinant VWF, preincubated with a radiolabelled monoclonal antibody (MAb) against VWF, is added to formalin-fixed platelets in the presence of increasing concentrations of ristocetin. In this method, VWF associated radioactivity bound to platelets is measured in the pellet. These methods have been widely used, with some minor modifications, to study biochemical characteristics of plasma VWF–GPIb interaction (Mohri et al, 1989; Matsushita & Sadler, 1995; Miura et al, 2000; Nakayama et al, 2002), or to test expressed recombinant VWF with various candidate mutations (Hilbert et al, 1995, 1998, 2002; Hillery et al, 1998; Facey et al, 1999; Ribba et al, 2001; Stepanian et al, 2003). However, these methods have some limitations for diagnostic laboratories, such as the need of patient's fresh platelets for RIPA or preparation of washed normal platelets, and licence for radioactivity handling.
We present here a new assay as an alternative for the study of the ristocetin-induced VWF binding to platelets. In the ELISA assay set up to measure VWF:RCo (Vanhoorelbeke et al, 2000) fresh or formaldehyde-fixed normal platelets are no longer required as the binding of VWF to a recombinant fragment of GPIb (His1–Val289) (rfGPIbα), purified as described by Schumpp (Schumpp-Vonach et al, 1995) is measured. This fragment also binds to VWF in the presence of botrocetin, with a comparable affinity as the proteolytic 42 kDa fragment of the purified human platelet GPIb/V/IX complex (Schumpp-Vonach et al, 1995). Furthermore, in contrast to the previously described platelet-based assay (Meulien et al, 1992), here rabbit anti-human VWF antibodies labelled with an enzyme are used rather than radioactivity. This test is able to readily distinguish type 1 from type 2 VWD samples (Vanhoorelbeke et al, 2002). However, so far, further differentiation of type 2 subtypes was not possible.
We report on the results obtained with this new test in 31 type 2B VWD patients with eight known mutations located in or close to the A1 disulphide loop, and we show that, this test can be used to discriminate between normal plasma and plasma of type 2B VWD patients.
Materials and methods
Normal and VWD patients plasma samples
Blood samples were collected by venepuncture in tubes containing 3·8% (0·129 mol/l) sodium citrate; PPP was obtained after double centrifugation at 2000 g for 15 min, and was stored at −80°C until use. Normal plasmas were obtained from 12 voluntary donors (EFS-Nord de France, Lille, France) with different ABO blood groups, to prepare the normal plasma pool (NPP). Plasma samples from 31 type 2B VWD patients from 15 families were studied. All patients displayed a positive RIPA at 0·5 mg/ml ristocetin concentration. Eight different mutations (H1268D, R1306W, R1306Q, R1306L, R1308C, V1316M, R1341Q and A1461V) in the exon 28 of VWF gene have been previously identified in these families (Meyer et al, 2001) (Table I). As a control of non-enhanced binding, seven type 2A VWD patients with three different mutations (R1315C, R1597W and L1639D) were also tested.
Table I. Biological description of the 2B VWD patients.
Mutation in pre-pro VWF
Platelets number (109/l)
Ratio VWF:RCo /VWF:Ag
% HMW multimers*
(a) Plasma VWF
(b) Platelet VWF
VWD, von Willebrand disease; VWF, von Willebrand factor; RCo, ristocetin co-factor; HMW, high-molecular weight; Ag, antigen.
*Percentages relative to a pool of normal plasmas.
†Ratio of OD of the patient sample when compared with OD of the NPP (a) or normal platelet lysate (b) inserted in the same assay.
Platelets were prepared from 15% ethylene diaminetetraacetic acid blood samples. PRP was obtained after centrifugation at 200 g for 15 min and the platelet number was counted. The PRP was then centrifuged at 3000 g for 10 min and the platelet pellet was stored at −80°C. Before use, the pellet was resuspended and lysed in phosphate-buffered saline (PBS) containing 0·5% Triton X-100 to adjust the platelet concentration to 3 × 109/ml and centrifuged at 10 000 g for 5 min. The VWF:Ag level in the supernatant was measured by ELISA (Mazurier et al, 1980). Platelet lysates from six normal subjects and 13 type 2B VWD patients from families 4, 7, 9, 10, 11, 12, 13 and 15 were studied (Table I).
Ristocetin-induced binding of VWF to rfGPIbα assay
A polystyrene plate (Maxisorp Nunc, Rochester, NY, USA) was coated, overnight at 4°C, with 100 μl of 5 μg anti-GPIbα MAb 2D4 in 50 mmol/l carbonate buffer pH 9·6. MAb 2D4 has a conformation-dependent epitope and has no inhibitory activity on ristocetin-induced human platelet agglutination (Vanhoorelbeke et al, 2000). The plate was then saturated with 3% milk powder (in 50 mmol/l PBS pH 7·4–250 μl/well) for 90 min at room temperature (RT), followed by a 2 h incubation at 37°C with a solution of 1 μg/ml of rfGPIbα in PBS, 0·1% Tween-20 (TBS) (100 μl/well). Meanwhile, 500 μl of either NPP or the test PPP sample – adjusted to 7 IU/dl VWF:Ag (in TBS) – were mixed with 10 μl of a 0·2 mg/ml dilution of anti-VWF-HRP (horse radish peroxidase-conjugated rabbit anti-human VWF polyclonal antibodies; Dako, Glostrup, Denmark). The non-inhibitory activity of the anti-VWF-HRP on the VWF binding has been preliminarily verified (data not shown). After a 1-h incubation at RT, 50 μl of the mixture and 50 μl of a serial ristocetin (Diagnostica Stago, Asnières, France) dilution (0–1 mg/ml at final concentration) were added to the wells coated with rfGPIbα for an incubation of 20 min at 37°C. After each incubation step, the plate was washed five times with TBS. The detection of captured ristocetin-VWF–anti-VWF HRP complexes on the rfGPIbα was done with ortho-phenylenediamine (3 mg/ml in 40 mmol/l sodium citrate – 10 mmol/l citric acid pH 5 and 0·03% H2O2– 100 μl/well) and the colouring reaction was stopped with H2SO4 (2 mol/l). The absorbance (OD) was read at 490 nm. The results of the VWF binding were expressed in OD as a function of ristocetin concentration. For evaluation of the type 2B VWD phenotype, the ratio of OD of the patient plasma sample over the OD of the NPP measured in the same assay (ratio OD sample/OD NPP) was calculated at the low ristocetin concentration of 0·2 mg/ml.
Ristocetin-induced binding of normal samples (plasma and platelets)
The binding of VWF from normal plasmas (n = 12) and from NPP (13 different experiments performed in a 18-month period) to the captured rfGPIbα in the presence of increasing concentrations of ristocetin was evaluated. Figure 1 shows the mean and SD of the results obtained. As expected, no binding of plasma VWF to rfGPIbα was observed in the absence of agonist and this binding was ristocetin-dose dependent.
When platelet lysates containing platelet VWF, from six normal individuals were tested, there was still no binding in the absence of ristocetin but the dose-dependent ristocetin-induced binding gave OD values that were higher than those obtained with NPP, as shown in Fig 1.
Binding of type 2B plasma or platelet VWF to rfGPIbα
Patients with the R1341Q mutation
The ability of plasma VWF from type 2B VWD patients to bind to rfGPIbα in the presence of increasing ristocetin concentrations was, for all the 14 patients (families 10 to 14) with the R1341Q mutation (Fig 2) the most informative; the graphics have been amplified in the area between 0 and 0·5 mg/ml of ristocetin. Binding of R1341Q-VWF from each patient was compared with that obtained with the NPP tested in the same experiment. Typical curves were obtained with significant enhanced binding at the lowest ristocetin concentrations with the most discriminant response at 0·2 mg/ml ristocetin. So, the ratio (OD sample/OD NPP) in the absence of ristocetin (for evaluation of spontaneous binding) or at 0·2 mg/ml ristocetin (for evaluation of 2B VWD phenotype) was calculated. Ratios for all samples tested, and the indicated mutation are given in Table I. In contrast to NPP or normal platelets, spontaneous binding to rfGPIbα may occur with patients sample, irrespective of the mutation and was without significant difference between the mutations, with a mean OD ratio of 1·64 (SD = 0·75) in 2B plasmas and 2·72 (SD = 1·29) in 2B platelets.
These 14 patients have been tested for the capacity of plasma VWF to bind to rfGPIbα (Fig 2A–D). In families 10 (Fig 2A) and 11 (Fig 2B), all relatives displayed increased binding at 0·2 mg/ml ristocetin; the OD ratio varied between 4 and 4·35 for family 10 and between 1·87 and 2·43 for family 11. In families 12 (Fig 2C) and 14 (Fig 2D), some variability in plasma VWF affinity was observed between the relatives, although, they shared the same molecular abnormality. Indeed, patient 12-1 shows only a slight increase at low ristocetin concentrations (ratio at 0·2 mg/ml = 1·37) whereas this ratio was 2·67 for patient 12-2. Patient 14-1 displayed a greater affinity for rfGPIbα (ratio at 0·2 mg/ml = 6·85) than his relative 14-3 (ratio only of 2·21). Patient 13-1 (Fig 2C), the only subject tested in this family, showed a typical binding curve with an OD ratio of 2·34 at 0·2 mg/ml ristocetin. On the whole, except for patient 12-1, the OD ratio at 0·2 mg/ml ristocetin was >1·8 for all the patients belonging to families 10 to 14 (Table I). Thus, we considered that a ratio above 1·8 might be indicative for a type 2B phenotype.
Platelet lysate samples. Six patients (from families 10 to 13) have been tested for the ability of their platelet VWF to bind to rfGPIbα. The increased binding at low ristocetin concentrations, typical of type 2B VWD, was clearly evident in all the platelet lysate samples tested (ratio at 0·2 mg/ml >2 – 2·06 to 5·26). It is noteworthy that, platelet lysate from patient 12-1, with a plasma VWF ratio of 1·37 at 0·2 mg/ml, displayed a significant increased platelet VWF binding at 0·2 mg/ml ristocetin, with a ratio of 2·65 (Fig 3).
Patients with mutations at position R1306
Three different mutations at position 1306 have been identified in the patients included in this study. Patients from families 2 and 3 harboured the R1306W mutation, whereas patients from families 4 and 5 displayed the R1306Q and R1306L mutations, respectively, (Table I). An enhanced binding of plasma VWF at 0·2 mg/ml ristocetin (ratio >2 – 2·09 to 4·61) was observed for all patients except for patient 3-1, with a ratio of 1·47.
Platelet lysate samples. Only platelets from three of the four relatives in family 4 were available for study. VWF in all platelet lysates showed an increased affinity with ratios of >4 (4·08–5·51) at 0·2 mg/ml ristocetin.
Patients with other mutations
The H1268D, R1308C, V1316M and A1461V mutations were reported in six families included in this study (Table I).
Plasma samples. Whereas, patient 7-1 with the R1308C mutation did not show an increased ratio at 0·2 mg/ml ristocetin (ratio = 1·12), the two other R1308C patients displayed values >2·7. For the two patients with the V1316M mutation, this value was >3 whereas for the patients with the A1461V mutation, it was 12·25 for 15-1 and 6·91 for 15-2 (Table I). It is noteworthy that, patient 15-1 also displayed the highest spontaneous binding (ratio in the absence of ristocetin = 4·76).
Platelet lysate samples. Platelet VWF from patients 7-1 (Fig 3), 9-1 and 15-1 exhibited the typical increase (ratio >2·5) at 0·2 mg/ml ristocetin.
Binding of type 2A plasma or platelet VWF to rfGPIbα
As expected, all the seven type 2A VWF plasmas tested did not show an increased OD ratio at 0·2 mg/ml ristocetin; furthermore, this ratio was always lower than 1 (from 0·68 to 0·98) and similar to that obtained in normal plasmas (from 0·71 to 1·02).
Negative correlation between the increased affinity of patient VWF for rfGPIbα and the platelet count
Platelet count varied from 47 × 109/l to 467 × 109/l (mean ± SD: 213 ± 92·109/l) in the 28 type 2B VWD patients studied (Table I); only five patients (patients 9-1, 10-1, 14-1, 14-2 and 15-1) had thrombocytopenia (<150 × 109/l). The plasma VWF from these five patients displayed high binding at 0·2 mg/ml ristocetin, with a ratio ranging from 3·63 to 12·25. VWF from patient 15-1, with the most severe thrombocytopenia (47 × 109/l) showed the highest binding to rfGPIbα (ratio at 0·2 mg/ml ristocetin = 12·25). It is noteworthy that the three patients aged ≤10 years (patients 8-1, 11-2 and 11-3), had the highest platelet counts; 467, 394 and 324 × 109/l respectively. The ratio at 0·2 mg/ml of plasma VWF was very similar in these three patients (between 2·28 and 2·72).
Furthermore, there was a significant negative correlation (r = −0·596; P = 0·016) between the platelet count and the OD ratio at 0·2 mg/ml ristocetin, in type 2B VWD patients >10 years in age.
Absence of correlation between the increased affinity of patient VWF for rfGPIbα and the percentage of HMW multimers
The percentage of HMW multimers above 10 mers was evaluated in samples from 24 patients (Table I). No correlation between the extent of the lack in HMW multimers and the enhanced affinity of VWF for rfGPIbα at 0·2 mg/ml ristocetin could be found (r = 0·353, P = 0·077).
Type 2B VWD refers to qualitative variants with increased affinity of VWF for platelet GPIb. These patients are generally characterised by an increased RIPA as well as an enhanced binding capacity of plasma VWF to platelets at low ristocetin concentrations. The RIPA uses PRP and has to be performed as soon as possible (within 2 h) after the sample is taken. Furthermore, RIPA does not discriminate between a plasma VWF hyperaffinity for GPIb (type 2B VWD) and a platelet GPIb hyperaffinity for VWF (pseudo-VWD). Therefore, evaluation of the ability of the plasma VWF to bind to platelets is usually performed to discriminate between these two abnormalities. The previously reported methods require radiolabelled anti-VWF antibodies and normal formaldehyde-fixed platelets (Meulien et al, 1992; Miura et al, 2000; Hilbert et al, 2002; Stepanian et al, 2003), which limits their straightforward applicability in most laboratories because of the need for a radioactivity licence and access to normal platelets for in vitro use.
We describe here a new method to measure ristocetin-induced binding of plasma or platelet VWF to GPIb that overcomes these limitations. We have previously shown (Vanhoorelbeke et al, 2000) that the VWF:RCo measurement can be performed using a recombinant fragment of GPIb that has been captured by an anti-GPIb antibody, and revelation of the VWF–GPIb complex by commercial HRP-labelled anti-VWF antibodies. We have also demonstrated (Vanhoorelbeke et al, 2002) that, with this test in combination with the determination of VWF antigen levels, one can readily distinguish type 1 from type 2 VWD. The present study has confirmed that, despite some modifications, normal plasma or platelet VWF binds to rfGPIbα in the presence of ristocetin and that this interaction is dose-dependent. This modified test can be used to identify patients with type 2B VWD mutations, where the ratio OD patient/OD NPP, determined at 0·2 mg/ml ristocetin and at a constant concentration of 0·70 IU/dl VWF:Ag, was found to be a very good indicator to identify type 2B VWD patients. Indeed, plasma VWF from 28/31 type 2B VWD patients with known mutations showed an enhanced OD ratio (>1·8) at 0·2 mg/ml ristocetin.
In plasma of type 2B VWD patients, however, the larger, more active VWF multimers are no longer present and this may confound the observations. Nevertheless, as the normal range of VWF multimers is present in the platelet granules (De Groot et al, 1989), the platelet VWF-binding capacity may be more straightforward. Indeed, VWF in the platelet lysates from all twelve patients tested showed an enhanced binding to rfGPIbα, including two of the three patients that were not identified when using their plasma. For the last patient, no such evidence could be obtained despite the fact that the results of all the other members of the family were unequivocal after re-evaluation. Our results further confirm that some variability in this ratio level may exist not only between patients with different mutations, but also between patients with the same mutation and even in patients from the same family. Such variability has been already reported with previous tests (Gaucher et al, 1995).
On the other hand, it was also noted that the two patients with the highest responses (ratio = 6·91 and 12·25), had only subnormal VWF:RCo/VWF:Ag ratios, which did not initially suggest a type 2 VWD, and had the A1461V mutation, which is located outside the A1 loop (Hilbert et al, 1995).
As it is known that type 2B VWD induces a loss of HMW multimers and mild thrombocytopenia, we evaluated whether a correlation exists between the OD ratio at 0·2 mg/ml and these two features. Indeed, plasma VWF reactivity for rfGP1bα in the presence of low doses of ristocetin may be relative to the importance of the lack in HMW multimers. So, in VWF with A1461V mutation there is only a mild decrease in HMW multimers as observed in patient 15-1. Consequently, the VWF still reacts strongly with platelets and rfGP1bα in the presence of low doses of ristocetin. So, although we did not find statistical correlation between the percentage of residual HMW multimers and the OD ratio at 0·2 mg/ml ristocetin, it is possible that the interpretation of our test may be less easy in patients harbouring more reactive point mutations, as the major lack of the HMW multimers does not enable a reaction with GPIb. In these cases, it will be useful to test the VWF from patient platelets. Effectively, platelet VWF from type 2B patients, similar to endothelial VWF, is normally multimerised and has an increased affinity for platelet GP1b, even in the absence of ristocetin (De Groot et al, 1989).
It has been previously documented in a large kindred (Mazurier et al, 1988) that platelet count is age-dependent, with the occurrence of thrombocytopenia only in adult affected members. This was confirmed in the present study by the absence of thrombocytopenia (platelet count >300 × 109/l) in three patients younger than 10 years. On the other hand, a relationship was found between OD at 0·2 mg/ml ristocetin and thrombocytopenia (platelet count <150 × 109/l) in patients older than 10 years from five families, where the more reactive type 2B plasmas were linked to more severe thrombocytopenia.
In conclusion, this study has extended the applicability of the test to enable the diagnosis of type 2B VWD by further optimisation. This test has several advantages over the classical methods as: (i) it can be performed on frozen plasma samples, without the need to obtain fresh PRP as required for a RIPA test; (ii) it is simple to perform as it does not use radiolabelled anti-VWF antibodies, and is therefore, accessible for laboratories without a licence to use radioactive material and (iii) it is sensitive, as 28/31 known type 2B patients showed typical plasma-binding curves with an enhanced OD ratio (>1·8) at 0·2 mg/ml ristocetin. Furthermore, it may be also applied in biochemical VWF interactions studies on rVWF expressed to reproduce various candidate VWD mutations. Figure 4 summarises the tests that need to be successively performed for the accurate diagnosis of types and subtypes of VWD.
We are very grateful to Drs A.Voyer and J.Diéval (Amiens), ME Briquel (Nancy), C. Boinot (Poitiers), K. Pouymayou (Marseille), L. Rugeri (Lyon) for providing VWD patient plasmas, and to S. Hermoire and D. Grenier for excellent technical assistance.