COAGULATION AND TRANSFUSION MEDICINE Original Article

The Evaluation of Factor VIII Binding Activity of von Willebrand Factor by Means of an ELISA Method Significance and Practical Implications

ALESSANDRA CASONATO, PhD, ELENA PONTARA, PhD, PATRIZIA ZERBINATI, PhD, ALESSANDRO ZUCCHETTO, MD, AND ANTONIO GIROLAMI, MD Downloaded from https://academic.oup.com/ajcp/article/109/3/347/1758043 by guest on 28 September 2021

One of the functions of von Willebrand factor (vWF) is to serve as chromogenic assay and our ELISA. A patient who was homozy­ a carrier of clotting factor VIII (FVIII). Deficiency of this function gous for the R53W mutation and had no FVIII binding capacity results in the von Willebrand disease (vWD) variant type 2N, according to the chromogenic method showed undetectable FVIII which resembles hemophilia A. We describe a new sandwich binding by ELISA. The remaining two patients, one who was enzyme-linked immunosorbent assay (ELISA) to study the abil­ homozygous for the R91Q mutation and one with compound het­ ity of vWF to bind exogenous recombinant FVIII (rFVIII), in erozygosity for the R91Q and R53W mutations, showed which anti-vWF-coated plates are incubated with plasma vWF, markedly decreased FVIII binding by the chromogenic method followed by exogenous FVIII and a peroxidase-coupled anti- and yielded ELISA values ranging from 4 to 8 U/dL. Therefore, FVIII antibody. Dose-response curves obtained using normal although the two methods produce qualitatively similar results, plasma vWF and purified normal vWF revealed a hyperbolic rela­ the ELISA method offers the advantage of allowing quantifica­ tionship between the optical density and the vWF concentration. tion of the FVIII binding function. FVIII binding was also ana­ The assay allows the quantification of FVIII binding with values lyzed in 20 patients with type 1 vWD; we found a decrease of expressed in U/dL; 100 U/dL was the amount present in normal FVIII binding that was proportionate to the decrease in vWF lev­ plasma. The sensitivity and specificity of the method are demon­ els, showing a normal FVIII binding activity/vWF molecule ratio. strated by its ability to measure binding levels as low as 1 to 2 We define the binding activity measured by this assay as U/dL and the fact that no FVIII binding was observed using vWF:FVIII binding activity and propose its use in the functional plasma known to contain less than 1 U/dL vWF. To verify the analysis of vWF. (Key words: Factor VIII; von Willebrand factor; accuracy of the assay, three patients with type 2N vWD with char­ FVIII binding activity; von Willebrand disease; Type 2N vWD) acterized vWF gene mutations were studied using an existing Am J Clin Pathol 1998;109:347-352.

Von Willebrand factor (vWF) is a high-molecular- includes vWF abnormalities characterized by defec­ weight glycoprotein that fulfills a crucial role in tive functions independent of multimer representa­ hemostasis by promoting platelet adhesion at the site tion. This group includes type 2N vWD, a variant of vascular injury1 and by serving as a carrier for clot­ previously called type Normandy vWD. Type 2N ting factor VIII (FVIII) in the plasma,2 thereby pro­ vWD is characterized by defective binding of vWF to tecting FVIII from proteolytic degradation.3'4 FVIII9-10 that is associated with normal platelet adhe­ Deficiency of vWF is responsible for the most fre­ sion. The phenotypic expression is characterized by quent inherited bleeding disorder, von Willebrand levels of FVIII that are substantially lower than the disease (vWD),5 in which quantitative defects (type 1 levels of vWF Although the condition is similar to and type 3) and molecular abnormalities (type 2) mild hemophilia A, its inheritance is not X-linked; have been described.6'7 According to the most recent nevertheless, misdiagnosis of hemophilia A has been proposed classification of vWD,8 the type 2 group described in type 2N vWD.11'12 Point mutations resulting in the amino acid substitutions Argl9Trp, Thr28Met, Arg53Trp, His54Gln and Arg91Gln have 13 16 From the University of Padua Medical School, Institute of Medical been reported in type 2N vWD. " Semeiotics, Padua, Italy. The FVIII-vWF interaction has been studied in This work was supported by grants from Ministero Universita' vitro using a chromogenic assay that estimates the Ricerca Scientifica e Tecnologica, Rome, Italy, and from Veneto exogenous FVIII binding to vWF immunoisolated Region, Venice, Italy. from plasma.4,5'8 We describe a new quantitative Manuscript received March 7,1997; revision accepted June 26,1997. Address reprint requests to Dr Casonato: Istituto di Semeiotica method that measures the binding of FVIII to Medica, Via Ospedale Civile 105, Padova, Italia. immobilized vWF.

347 348 COAGULATION AND TRANSFUSION MEDICINE Original Article

MATERIALS AND METHODS RESULTS

Patients and healthy volunteers were studied. Demonstration of the Specific Binding Informed consent was obtained in accordance with of Recombinant FVIII to Immobilized vWF the declaration of Helsinki. Fifty healthy volunteers (30 women and 20 men), 21 to 65 years of age, who A sandwich immunoassay was developed to had not taken medication for at least 2 weeks before quantify the capability of plasma vWF to bind to the study served as controls. Nineteen patients with exogenous rFVIII. As described in the "Materials and type 1 subtype platelet-normal, 1 patient with type 1 Methods" section, the assay used microtiter plates

platelet-low, and 1 patient with type 3 vWD were coated with anti-vWF antibody, which were incu­ Downloaded from https://academic.oup.com/ajcp/article/109/3/347/1758043 by guest on 28 September 2021 studied. Blood was drawn from the antecubital vein bated with plasma samples followed by rFVIII and and anticoagulated using 3.8% sodium citrate (1:10 horseradish peroxidase-conjugated anti-FVIII anti­ vol/vol). Platelet-poor plasma was prepared and body. Figure 1 shows the dose-response curves bleeding time was measured as previously obtained with serial dilutions of purified vWF and described.17 Assays measuring vWF ristocetin cofac- vWF present in NP at a constant concentration of tor activity (vWF:RCo), vWF antigen (vWF:Ag), and exogenous rFVIII (1.0 U/mL). Results showed that FVIII activity (FVIILC) were previously reported.18-20 the binding of rFVIII was a function of the amount of The vWF was purified from human cryoprecipitates according to the method of De Marco and Shapiro.21

Binding Assay of Recombinant FVIII to Immobilized vWF Purified normal von Willebrand factor Microtiter wells in 96-well plates were coated 1,500 Normal pooled plasma overnight at 4°C with 200 uL of polyclonal anti-vWF Type 3 von Willebrand disease antibody (Dako, Amsterdam, the Netherlands), 1.5 ug/mL in 0.05 mol/L sodium carbonate-bicarbonate buffer, pH 9.6. After extensive washing, the wells were incubated with phosphate-buffered saline (PBS) en 1,000 containing 2% bovine serum albumin (BSA) for 1 c hour at room temperature. The following steps were D then performed at room temperature. After washing, 200 uL of selected plasma dilutions, prepared in PBS 8 containing 0.05% polysorbate (Tween) and 2% BSA, were incubated in the wells for 1 hour. Endogenous FVIII was then removed from the immobilized vWF by incubation with buffered 0.4 mol/L calcium chlo­ ride for 30 minutes. After washing selected dilutions of recombinant FVIII (rFVIII; Baxter, Deerfield, 111), prepared in a PBS-Tween-BSA mixture were added and incubated for different times. After washing, the 1:100 1:400 1:800 1:1,600 bound rFVIII molecule was detected by incubation with horseradish peroxidase-conjugated polyclonal Plasma Dilution anti-FVIII antibodies (Enzymes Research, South Bend, FIG 1. Binding of recombinant factor VIII (rFVIII) to immobilized von Ind). The color was developed by addition of ortho- Willebrand factor (vWF) from normal pooled plasma (NP), purified normal vWF, and type 3 von Willebrand disease (vWD) plasma. The phenylenediamine dihydrochloride. The values were dose-response curves were obtained by incubating immobilized vWF plotted against plasma vWF dilutions and were with a constant amount of rFVIII (1 U/mL) for 30 minutes at 22°C. expressed as U/dL, with 100 U/dL the value The bound rFVIII was detected and quantified using horseradish per­ obtained with a 1:25 dilution of normal pooled oxidase-conjugated anti-FVIII antibody. The color was developed by plasma (NP). The ratio of FVIII binding activity to addition of o-phenylenediamine dihydrochloride. The NP was serially diluted starting from 1:25, and purified normal vWF was serially vWF:Ag was also calculated to normalize the binding diluted starting at a concentration of 0.4 |ig/mL, which corresponds to function to the vWF levels. the amount of vWF present in a 1:25 dilution of NP.

AJCP • March 1W8 CASONATO ET AL 349 FVIII Binding to v\ Evaluated by ELISA immobilized vWF molecule and that the decrease in Binding of FVIII to Plasma vWFFrom Healthy FVIII binding had a hyperbolic relationship to the Persons and Patients With Type 1 vWD decrease in vWF. The specificity of rFVIII binding to vWF was demonstrated by the dose-response curve Fifty healthy subjects were studied to define a nor­ obtained with normal purified vWF, which showed a mal range of FVIII-vWF binding; three dilutions (1:25, relationship similar to that obtained using NP at sim­ 1:50, 1:100) were performed for each sample. ilar concentrations (see Fig 1). As a further test for Moreover, vWF:Ag was also evaluated by using an possible nonspecific binding of rFVIII to plasma pro­ ELISA method that uses the same coated anti-vWF teins other than vWF, plasma from a patient with antibody at a similar concentration. For the FVIII type 3 vWD and vWF levels known to be below to 1.0 binding assay, the standard calibration curve was Downloaded from https://academic.oup.com/ajcp/article/109/3/347/1758043 by guest on 28 September 2021 U/dL was assayed. As shown in Figure 1, no color was developed using this plasma; furthermore, non­ specific binding to the coated anti-vWF antibody was not observed when plasma was replaced with dilut­ ing buffer (data not shown). Experiments performed with serial dilutions of rFVIII (from 1.5 to 0.125 U/mL) and a constant amount of coated vWF demonstrated a direct linear relationship between the optical density and log of the rFVIII concentration (Fig 2, A). When the amount of the immobilized plasma vWF was reduced (NP diluted from 1:25 to 1:100), the slope of the dose-response line decreased substantially. Similarly, the dose-response curve obtained with a constant concentration of rFVIII demonstrated a progressive decrease in optical density with consec­ 1:50 1:200 1:800 rFVIII (U/mL) Plasma Dilution utive dilutions of vWF, and the slope of the line also decreased when rFVIII was reduced from 1.5 U/mL FIG 2. A, Relationship between optical density and serial dilutions of recombinant factor VIII (rFVIII) measured using a constant to 0.125 U/mL (Fig 2, B). These findings confirmed amount of von Willebrand factor (vWF), with each assay per­ that the amount of immobilized vWF and the con­ formed using the following dilutions of normal pooled plasma centration of the exogenous rFVIII affect the binding (NP): 1:25, 1:50, 1:100, and 1:200. Experimental conditions are assay. Subsequent dose-response curves were con­ described in the legend for Figure 1. B, Relationship between opti­ structed by using serial dilutions of immobilized cal density and the amount of vWF in serial dilutions of NP, with vWF, starting from 1:25, and a constant concentra­ each assay performed using the following concentrations of rFVIII: 1.5 U/mL, 0.75 U/mL, 0.25 U/mL, and 0.125 U/mL. The rFVIII tion of rFVIII (1.0 U/mL). incubation step was carried out for 1 hour at 22°C.

Experimental Conditions Affecting the Binding of Exogenous FVIII

The effects of temperature and time of the FVIII incubation step on vWF-FVIII binding were studied. At 22°C, prolonging the rFVIII incubation time from 5 to 60 minutes produced a slight increase in the optical density (Fig 3, A). When rFVIII was incu­ bated at 37°C, optical density values increased slightly compared with those obtained at 22°C, but i i i i i i 1 i I i i i i i i i 1:25 1:100 1:400 1:1,600 1:25 1:100 1:400 1:1,600 longer incubation times increased the background 1:50 1:200 1:800 1:3,200 1:50 1:200 1:800 1:3,200 color (Fig 3, B). Moreover, an inverse relationship Plasma Dilution between the rFVIII concentration and incubation time was found; shorter incubation times were suffi­ FlG 3. Dose-response curves of factor VIII (FVIII) binding to immo­ bilized vWF obtained by incubating with recombinant FVIII (1 cient when higher concentrations of rFVIII were U/mL) at 22°C (A) or 37°C (B) for 60 minutes, 30 minutes, 15 min­ used (data not shown). utes, 10 minutes, or 5 minutes.

Vol.: • No. 3 350 COAGULATION AND TRANSFUSION MEDICINE Original Article obtained by progressive dilutions of NP (from 1:25 to homozygous for the R53W mutation; patient 2 was 1:3,200); the measured values were expressed in homozygous for the R91Q mutation; and patient 3 U/dL, with 100 U/dL the value obtained with 1:25 had compound heterozygosity for the R91Q and dilution of NP. The normal mean ± SD FVIII binding R53W mutations. When the ELISA method was used, value was found to be 95.6 ± 14.75 U/dL (range, the sample from patient 1 showed undetectable FVIII 65-130 U/dL; mean ± 2 SD). The ratio of FVIII bind­ binding (values below 1 U/dL), while the samples ing activity to the level of vWF:Ag ranged from 0.6 to from patients 2 and 3 showed similar values, ranging 1.3. In the type 1 platelet-normal vWD samples tested, between 4 and 8 U/dL. When the existing chro- the FVIII binding activity seemed to be decreased in mogenic method was used, the sample from patient 1

proportion to the decrease in vWF:Ag (44.9 ± 8.9 was classified as having absent FVIII binding activity, Downloaded from https://academic.oup.com/ajcp/article/109/3/347/1758043 by guest on 28 September 2021 U/dL and 45.6 ±9.9 U/dL, respectively). Hence, the and the samples from patients 2 and 3 were classified FVIII binding ratio was within the normal range as having markedly decreased binding. (mean ± SD, 1.006 ± 0.16) in these patients. The lowest FVIII binding capacity value we were able to detect Identification of a Family With Type 2 N vWD was 2.3 U/dL, observed in a patient with type 1 platelet-low vWD who had a level of vWF:Ag of 3 Two subjects from the same kindred, previously U/dL; this observation indicated that the assay is classified as having type 1 vWD, showed a decrease quite sensitive. in FVIII binding activity that was more pronounced than the decrease in vWF:Ag (Table 2). The FVIII FVIII Binding in Patients With Recognized Genetic binding capacity ranged from 4 to 12 U/dL in com­ Mutations in the FVIII Binding Domain ofvWF parison with the vWF:Ag values, which ranged from 40 to 48.5 U/dL. The FVIII binding ratio was 0.15 to To verify the efficacy of the method, 3 samples 0.3, while the FVIII:C/vWF:Ag ratio ranged from 0.6 from patients with known defective FVIII binding of to 0.7, vs 0.8 to 1.2 for healthy subjects. Hence, the vWF and a genetic defect in the FVIII binding domain FVIII binding capacity of vWF in vitro was more of vWF (supplied C. Mazurier, PhD, Lille, France), abnormal than predicted by the relative decrease in were tested using our ELISA (Table 1). Patient 1 was FVIILC compared with vWF. These patients seemed

TABLE 1. COMPARATIVE EVALUATION OF FVIII BINDING ACTIVITY OF vWF*

FVIII Binding

Patients Mutation Chromogenic Method ELISA (U/dL)

1 Homozygous R53W Absent <1 2 Homozygous R91Q Markedly decreased 4-8 3 Compound heterozygous R91Q-R53W Markedly decreased 4-8 Healthy control subjects Present 65-130

FVIII = factor VIII; vWF - von Willebrand factor; ELISA = enzyme-linked immunosorbent assay. 'Measured by the chromogenic and ELISA methods in three patients with a characterized mutation in the FVIII binding domain of vWF.

TABLE 2. HEMOSTATIC FINDINGS IN THE FAMILY WITH ABNORMAL FVIII BINDING TO vWF

Blood BT PTT FVIILC vWF:Ag vWF:RCo FVIII Binding FVIII Binding! Subject Group (minis) (s) (U/dL) (U/dL) (UldL) (UldL) vWF:Ag Ratio

Proband B 3/25 47.5 24.5 40.0 48.7 8.7 0.21 Sister B 3/13 45.0 34.5 48.5 49.2 8.3 0.17 Daughter B 2/38 37.0 55.0 33.1 38.0 43.7 1.3 Son O 3/30 35.5 76.0 29.3 21.9 35.5 1.2 Normal range <4 30-40 60-160 60-160 60-130 65-130 0.6-1.3

FVIII = factor VIII; vWF = von Willebrand factor; BT = bleeding time; PTT = partial thromboplastin time; FVIIFC = FVIII activity; vWF:Ag = vWF antigen; vWF:RCo = vWF ristocetin cofactor activity.

A|CP • March 1998 CASONATO ET AL 351 FVI1I Binding to v\ Evaluated by ELISA to simultaneously show a decrease in vWF levels and vWF:FVIIIBA values ranging from 4 to 8 U/dL in a defective FVIII-vWF binding. The sons of one of these patient who was homozygous for the R91Q muta­ patients showed a decrease in vWF levels associated tion and in a patient with compound heterozygos­ with normal FVIII values (see Table 2), indicating nor­ ity for the R53W and R91Q mutations, both of mal FVIII binding activity. whom were previously described to have markedly decreased FVIII binding. In the patients with type 1 platelet normal vWD, the vWF:FVIIIBA values and DISCUSSION the vWF concentration were decreased proportion­ We describe a new ELISA method to measure the ately. The normalization of vWF:FVIIIBA to vWF:Ag, performed to compare their relative repre­ FVTII carrier capacity of vWF. The method allows a quan­ Downloaded from https://academic.oup.com/ajcp/article/109/3/347/1758043 by guest on 28 September 2021 titative evaluation of this function and seems to be quite sentation, demonstrated ratios between 0.6 and 1.3 specific and sensitive, detecting FVIII binding values in all healthy subjects and patients with type 1 ranging from 1 to 100 U/dL. The assay values seem to vWD, except two subjects with vWD who had a sig­ reflect the plasma vWF concentration and its FVIII carrier nificantly decreased ratio (between 0.15 and 0.3). capacity. We propose to define this parameter as These two subjects, from the same kindred, had vWF:FVIII binding activity (vWF:FVIIIBA). Currently, previously been classified as having type 1 vWD only vWF:RCo is used in vitro to study vWF function; based on their simultaneous decrease in vWF and vWF:RCo measures the adhesive function of vWF and is FVIII. Despite the mild relative decrease in plasma related to large multimer representation. Because another FVIII compared with vWF (FVIII:C/vWF:Ag ratios key function of vWF is to bind to FVIII and protect it ranging between 0.61 and 0.71), vWF:FVIIIBA from proteases, we propose that vWF:FVIHBA should be seemed to be substantially reduced, with values included in the characterization of vWF function. between 4 U/dL and 12 U/dL. This observation suggests that a defective FVIII-vWF binding capac­ Abnormality in the ability of vWF to bind FVIII is ity also may be present in patients with FVIII that is responsible for a recently described vWD variant not substantially more reduced than vWF. (type 2N, formerly type Normandy) characterized by Therefore, multiple factors may contribute in vivo a decreased capacity or total inability of vWF to bind to modulate the phenotypic expression of FVIII FVIII. This abnormal binding capacity was identified binding defects. In addition, the subjects in this kin­ by using a chromogenic method that uses blood coag­ 9 10 22 dred seemed to exhibit a quantitative reduction ulation factors IXa and X. ' - However, this method defect in vWF and defective vWF-FVIII binding as is only qualitative or semiquantitative and defines previously described.23"26 The defect was associ­ FVIII binding capacity as absent, markedly decreased ated with compound heterozygosity for R91Q and or moderately decreased. In contrast, the method we AC 2680 in exon 1812-27 (C. Mazurier, PhD, personal developed allows a quantitative evaluation of this communication, February 17, 1997). The availability function, calculated against a standard calibration of this simple ELISA method will allow the study of curve that is obtained by progressive dilution of nor­ the FVIII binding function of vWF in all patients mal plasma, and expresses the binding values in with quantitative defects, regardless of the repre­ U/dL. It can estimate the binding of FVIII to vWF at sentation of FVIII in relation to vWF. values lower than 1.5 U/dL. The specificity of the method is indicated by the fact that purified normal We established a new method to study and vWF shows a dose-response curve similar to that of quantify the FVIII carrier function of vWF. In con­ normal plasma. In addition, plasma from a patient trast to the previous cumbersome chromogenic with type 3 vWD shows vWF:FVIIIBA similar to that method, the present method is simple and sensi­ obtained with the buffer control, indicating the tive, features that should favor the measurement absence of nonspecific binding of other plasma pro­ of FVIII binding in a large cohort of patients with teins to rFVIII. vWD. We hope that application of this new We compared our results with those obtained by method will facilitate the evaluation of FVIII bind­ the chromogenic method in patients with type 2N ing activity of vWF in patients with vWD or hemo­ vWD and well-characterized vWF gene mutations. philia A so a correct diagnosis of type 2N vWD We observed undetectable vWF:FVIIIBA values (< 1 may be easily determined. U/dL) in a patient homozygous for the R53W Acknowledgments: We express gratitude to Claudine Mazurier, mutation who was previously found to have absent PhD, and her coworkers for providing plasma samples from binding by chromogenic assay; we also observed patients with type 2N vWD and for performing the genetic studies.

Vol.1 •No. 3 352 COAGULATION AND TRANSFUSION MEDICINE Original Article

REFERENCES 16. Kroner PA, Foster PA, Fash SA, et al. The defective interaction between von Willebrand factor and factor VIII in a patient 1. Sakariassen KS, Bolhuis PM, Sixma JJ. Human blood platelets with type 1 von Willebrand disease is caused by substitution adhesion to artery subendothelium is mediated by factor of Argl9 and His54 in mature von Willebrand factor. Blood. Vlll-von Willebrand factor bound to subendothelium. 1996;87:1013-1021. Nature. 1979;270:636-638. 17. Fabris F, Casonato A, Del Ben MG, et al. Abnormalities of von 2. Foster PA, Fulcher CA, Marti T, et al. A major factor VIII bind­ Willebrand factor in myeloproliferative disease: a relation­ ing domain resides within the amino-terminal 272 amino ship with bleeding diathesis. Br J Haematol. 1986;63:75-83. acid residues of von Willebrand factor. / Biol Chem. 18. Casonato A, Fabris F, Vicariotto M, et al. The evaluation of fac­ 1987;262:8443-8446. tor VIII antigen by means of a simple slide test. Am J Clin 3. Fay PJ, Coumans JV, Walker FJ. Von Willebrand factor mediates Pathol. 1985;29:309-316. protection of factor VIII from activated protein C-catalyzed inactivation. / Biol Chem. 1991;266:2172-2177. 19. Casonato A, Sartori MT, Pontara E, et al. Impaired release of tis­ sue plasminogen activator (t-PA) following DDAVP infusion

4. Koedam KA, Hamer RJ, Beeser-Visser NH, et al. The effect of Downloaded from https://academic.oup.com/ajcp/article/109/3/347/1758043 by guest on 28 September 2021 von Willebrand factor on activation of factor VIII by Xa. Eur ] in von Willebrand's disease with low platelet von Biochem. 1990;189:229-234. Willebrand factor content. Blood Coagul Fibrinolysis. 5. Rodeghiero F, Castaman G, Dini E. Epidemiological investiga­ 1992;3:149-153. tion of the prevalence of von Willebrand's disease. Blood. 20. Girolami A, Molaro G, Lazzarin M, et al. A new congenital 1987;69:454-459. haemorrhagic condition due to the presence of abnormal fac­ 6. Ruggeri ZM, Zimmerman TS. von Willebrand factor and von tor X (factor X Friuli): studv of a large kindred. Br j Haematol. Willebrand disease. Blood. 1987;70:895-904. 1970;19:179-192. 7. Ruggeri ZM, Pareti FI, Mannucci PM, et al. Heightened interac­ 21. De Marco L, Shapiro SS. Properties of human asialo factor VIII: tion between platelets and factor Vlll/von Willebrand factor a ristocetin-independent aggregating aspect. / Clin Invest. in a new subtype of von Willebrand's disease. N Engl j Med. 1981;68:321-328. 1980;302:1047-1051. 22. Lopez-Fernandez MF, Bianco-Lopez M, Castineira MP, et al. 8. Sadler JE. A revised classification of von Willebrand disease. Further evidence for recessive inheritance of von Willebrand Thromb Haemsost. 1994;71:520-525. disease with abnormal binding of von Willebrand factor VIII. 9. Mazurier C, Dieval ], Jorieux S, et al. A new von Willebrand fac­ Am J Hematol. 1992;40:20-27. tor (vWF) defect in a patient with factor VIII (FVIII) defi­ 23. Siguret V, Lavergne JM, Cherel G, et al. A novel case of com­ ciency but with normal levels and multimeric pattern of both plasma and platelet vWF: characterization of abnormal pound heterozygosity with 'Normandy ' type I von vWF/FVIH interaction. Blood. 1990;75:20-26. Willebrand disease (vWD): direct demonstration of the seg­ 10. Nishino M, Girma J-P, Rothschild C, et al. New variant von regation of one allele with a defective expression at the Willebrand disease with defective binding to factor VIII. mRNA level causing type I vWD. Hum Genet. Blood. 1989;74:1591-1599. 1993;93:95-102. 11. Mazurier C, Christine G, Jorieux S, et al. Evidence for a von 24. Peerlinck K, Eikenboom JC, Ploos Van Amstel HK, et al. A Willebrand factor defect in factor VIII binding in three mem­ patient with von Willebrand's disease characterized by a bers of a family previously misdiagnosed mild haemophilia compound heterozygosity for a substitution of Arg 854 by A and haemophilia A carriers: consequence for therapy and Gin in the putative factor-VIII-binding domain of von genetic counselling. Br j Haematol. 1990;76:372-379. Willebrand factor (vWf) on one allele and very low levels of 12. Schneppenheim R, Budde U, Krey S, et al. Results of a screen­ mRNA from the second vWF allele. Br } Haematol. ing for von Willebrand disease type 2N in patients with sus­ 1992;80:358-363. pected haemophilia A or von Willebrand disease type 1. 25. Kroner PA, Foster PA, Fash SA, et al. The defective interaction Thromb Haemost. 1996;76:598-602. between von Willebrand factor and factor VIII in a patient 13. Gaucher C, Jorieux S, Mercier B, et al. The Normandy variant of with type 1 von Willebrand disease is caused by substitution von Willebrand's disease: characterization of a point muta­ of Argl9 and His54 in mature von Willebrand factor. Blood. tion in the von Willebrand factor gene. Blood. 1996;87:1013-1021. 1991;77:1937-1941. 26. Mazurier C, Meyer D. Factor VIII binding assay of von 14. Cacheris PM, Nicholas WC, Ginsburg D. Molecular characteri­ zation of a unique von Willebrand disease variant: a novel Willebrand factor and the diagnosis of type 2N von mutation affecting von Willebrand factor/factor VIII interac­ Willebrand disease: results of an international survey. tion. / Biol Chem. 1991;266:13499-13502. Thromb Haemost. 1996;76:270-274. 15. Jorieux S, Tuley EA, Gaucher C, et al. The mutation Arg(53) Trp 27. Zhang ZP, Blomback M, Egberg N, et al. A single cytosine dele­ causes von Willebrand disease Normandy by abolishing tion in exon 18 of the von Willebrand factor gene is the most binding to factor VIII: studies with recombinant von common mutation in Swedish vWD type III patients. Hum Willebrand factor. Blood. 1992;79:563-567. Mol Genet. 1992;1:767-768.

AJCP • March 1998