Applications of immunoperoxidase techniques in specificity testing of monoclonal antibodies (Mabs) against von Willebrand factor (vWf)

Cltntca Chimica Actu, 174 (1988) 65-82 Elsf%ier CCA 04144

A~~~cat~ons of i~un~~~ro~dase techniques in specificity testing of ~o~oclonal antibodies (Mabs) against von Willebrand factor (vWf) 3.Ingerslev

a* A. Bulch ‘, N.P.H.

MPrller b, S. Stenbjerg

a and J. Zeuthen ’

a Department of Clinical immunology, Untverstty Hospiial Aarhur, ’ Instttute of Medical Microbiology, University of Aarhur, and ’ Novo BioLabs, Now industv A/S, Bagsvaerd (Denmark) (Received 9 September 1987; revision received 12 Flecember 1987; accepted after revision 27 &uuq 1988) Key words: Von Wiilebrand factor; Von Willebrand’s disease; Multimeric sizing; Fast protein liquid chromatography (FPLC)

The present complication describes the production of a new series of mu&e Mabs against von WiIlebrand factor (vWf) in which specificity was tested using i~unopero~dase techniques. Seven Mabs showed specific reactivity with native and disaggregated vWf, whereas no binding was found to material from patients with severe homozygous (or doubly heterozygous) von Willebrand’s disease (vWd) or factor VIII coagulant antigen (VIII: Ag). These Mabs are thought to carry separate specificities as only slight or no competitive activity was detected. Four Mabs partially inhibited the ristocetin-induced platelet a~lutination and three interacted with ~Wf-bind~g to type I collagen. All antibodies bound to the complete range of vWf multimers of normal plasma. Excellent binding and detection properties of Mabs were found in asymmetrical two-site enzyme linked immunosorbent assays (ELISA) for quantitation of vWf antigen (vWf :Ag). One particular antibody (Mab vWf-33) discriminated vWf material from a number of subtype II vWd plasmas tested.

The van Willebrand factor (vWf) is a large plasma protein (M, 12-20 x lo6 Da) which is synthesized in endothelial cells fl] under control of an autosomal gene

Correspondence to: Dr. J. fngerstev, Department DK-8000 Aarhus C. Denmark. ~-8~8~/88~~~~.~~

of Clinical Chemistry, University Hospital Aarhus,

Q 1988 Efsevier Science Publishers B.V. (Biomedical Division)

66

located on chromosome no. 12 [2,3]. During recent years, a number of monoclonal antibodies (Mabs) against vWf have been reported [4-91. These were important tools in the characterization of specific adhesive vWf domains, including Mabs specific for the part of vWf polypeptide responsible for platelet agglutination in the presence of the inducer antibiotic ristocetin [8,10,11] effected through vWf adhesion to platelet glycoprotein Ib, a second epitope of vWf binding to the glycoprotein IIb/IIIa complex of the activated platelet [12-141, and a site of vWf binding to subendothelium [8,9], to collagens [15], to heparin [16], and to factor VIII [17]. Practical applications of Mabs in immunoradiometric assays (IRMA) [18-201 and enzyme-linked imrnunosorbent assays (ELISA) [21-231 for quantitation of vWf have been reported, the latter used in screening of expressed vWf from transfected cells during recent molecular cloning of vWf [3,24]. Mabs have also have been useful reagents in histological detection of vWf [25]. A registry of Mabs against vWf including some major findings was recently reported [26]. In the present report we describe the production of a series of seven Mabs against vWf and the characterization of their particular specificities by extensive use of immunoperoxidase techniques. Demonstration of Mab binding to native plasma vWf separated by fast protein liquid chromatography (FPLC), use of Mabs in visualization of vWf oligomers, and Mab inhibitory activity against ristocetin induced platelet agglutination and collagen binding of vWf are reported. All Mabs were cross-tested in two-site ELISA to evaluate the individual binding and detection characteristics. The development of a two-site Mab based ELISA seemingly discriminating type II von Willebrand’s disease material is described. Materials and methods Preparation of v Wf antigen A commercial factor VIII (FVIII) concentrate was fractionated by Biogel A-5M chromatography using calcium-free Tris buffer 0.05 mol/l (Tris 6.05 g/l adjusted with HC10.1 mol/l), pH 7.5, containing Aprotinin 10 KIE/ml (Bayer, Leverkusen, FRG). The column (2.5 x 73 cm) was kept at 4°C and the flow was 30 ml/h. Fractions corresponding to the void volume were pooled and concentrated on a Diaflo PM10 membrane (Amicon Corporation, Danvers, MA, USA). The concentrate now yielded 1 mg/ml of protein assuming an E2*’ “m (I%, 1 cm) of 5.0. Three BALB/c mice were injected i.p. with 150 ~1 of a mixture of equal volumes of FVIII/vWf containing material and complete Freund’s adjuvant and at 2-wk intervals given four subcutaneous injections each of 50 ~1 of the same material. Three weeks after the last subcutaneous dose, an intravenous injection of either 25 or 100 ~1 of mixed antigen and adjuvant as above was injected, and the spleen was removed after 4 days. Production of hybridomas Spleen cells were fused with X63-Ag8-6.5.3 non-producing myeloma cells [27] using polyethylene glycol (PEG; mol wt 1500) and selected in 96-well plates in hypoxanthine-aminopterine-thymidine (HAT) medium [28,29]. Screening of super-

natans was performed in an ELISA, in which the FVIII concentrate used for immunization was applied to the solid phase and binding of mouse immunoglobulin was disclosed by use of a peroxidase-labelled rabbit antimouse immunoglobulin. Eight clones exhibiting binding were stabilized by recloning and studied further. Production of ascites Cells from each of the hybridomas were injected i.p. in pristane-primed BALB/c mice. Ascitic fluid was collected lo-15 days after inoculation and immunoglobulin obtained by precipitation at ambient temperature with saturated ammonia sulphate at 50% or by protein A-Sepharose Cl-4B purification (Pharmacia, Uppsala, Sweden), carefully following manufacturers recommendations. Assuming a E *” (1 cm, 1%) of 13 for mouse immunoglobulin, the yield of immunogobulin from ascitic fluid was in the region of S-6 mg/ml as determined by spectrophotometry. Isotype determination of immunoglobulins was performed by Ouchterlony diffusion test using specific antisera for murine isotypes (Serotec, UK). All Mabs were conjugated with peroxidase by the periodate method [30]. Polycional antibodies Rabbit anti-human vWf : Ag from Dakopatts (Copenhagen, Denmark) was prepared as F(ab’), fragments and conjugated with horse-radish peroxidase (HRP) as previously reported [31]. In a few experiments the commercial HRP-conjugate of undigested IgG-fraction of the same antibody was used (Dakopatts, code P-226). Rabbit antibodies against human fibrinogen, fibronectin, IgM, and ~~-macroglobulin, peroxidase conjugates of rabbit anti-mouse i~unoglobulins, biotinylated rabbit-anti-mouse immunoglobulin, avidin-peroxidase and peroxidase-conjugated swine-anti-rabbit immunoglobulin also were from Dakopatts. Biotinylated rabbit-antimouse immunoglobulins and avidin-peroxidase were obtained from the same source. Screening assay for Mab binding to native plasma v Wf The ELISA was as previously reported [31], but here vWf : Ag retained by the rabbit antibody (severe type vWd plasma serving as control) was detected by a two-step procedure using anti-vWf : Ag Mab followed by HRP-conjugated rabbit-anti-mouse immunoglobulin. Displacement studies by sequential competitive ELISA One hundred microlitres of normal plasma diluted 1: 10 in sample buffer (Na,HPO, 9.149 g, KH,PO, 2.078 g, and NaCl 4 g in 1 litre of sterile water) (PBS) containing Tween-20 1 g/l (PBS-Tween) and bovine albumin (Sigma A-7030) 10 g/1 at pH 7.2 was pipetted into microtitre plates previously coated with polyclonal anti-vWf antibody. After completion of binding, and three washes with washing buffer (NaH,PO, 0.345 g, Na,HPO,, 12 Hz0 2.680 g. NaCl 8.747, and Tween-20 1 g/l), several dilutions of Mabs were incubated in the ELISA-plate for 2 h at room temperature, and, following three washes as above, remaining free vWf : Ag epitopes were quantitated by HRP-conjugate of a different Mab or the F(ab’),-HRP

68

conjugate of the polyclonal antibody. This assay identified shared antigenic sites recognized by individual Mabs and the polyclonal antibody. Mab binding to von Willebrand factor multimers The principle of the multimeric analysis was reported before 1323.In the present

study, transfer of protein on to nitrocellulose filters was effected either through passive diffusion or by electrotransfer as originally described. Lanes of nitrocellulose blots of vWf multimers of controls’ or patients’ plasma were separately incubated with individual Mabs. In some cases, a two- or three-step procedure was used in which bound Mabs were detected by either HRP-conjugated anti-mouse immunoglobulins or by biotinylated anti-mouse immunoglobulins followed by avidin-peroxidase (Dakopatts). In others, HRP-conjugates of Mabs were employed for direct detection of oligomers. Mab inhibition of v Wf-binding to collagen

The analysis followed guidelines recently reported [33]. Collagen (Helena Labs., Beaumont, TX, USA) was diluted to 30 pg,/ml in a 0.1 mol/l acetate buffer (C,O,Na, 3 H,O 9.23 g, acetic acid, 1.93 g in 1 litre), pH 5.0, and used for coating of microtitre plates overnight in the cold. After emptying of the plates, bound collagen was fixed to wells by heating the microtitre plate bottom with a hair drier. Fibrillar a~~gement was seen in the inverted microscope. 1: 10 diluted normal plasma samples in PBS-Tween were preincubated for 1 h at room temperature with serial dilutions of Mab material using anti-TNP [34] as negative control. After two washes with isotonic saline (NaC19 g/l) containing Tween-20 0.05%, 100 ~1 of each preincubated sample were applied in duplicate, and incubated at room temperature for 2 h. After three washings as above, a HRP-conjugate of another and non-crossreactive Mab, appropriately diluted in PBS, pH 7.2, with Tween-20 0.5 g/l was incubated for 2 h. After three washings as before, enzyme activity was quantitated using a mixture of o-phenylendia~ne (4 tablets, 2 mg each) (Dakopatts) in 12 ml of citrate-phosphate buffer 100 mmol/l (citric acid, H,O 7.30 g, Na,HPO,, 12 H,O 23.88 g/l) pH 5.0, and 40 ~1 of urea-peroxide solution (1 tablet, 0.35 g in 10 ml of sterile water) (BDH, Poole, UK), and the reaction was stopped by addition of 2 mol/l sulphuric acid. The absorption was read at 492 nm and inhibition of collagen binding was expressed as percentage of displacement compared with the anti-TNP control. Mab inhibition of the ristocetin cofactor

Mab interaction with the ristocetin induced platelet agglutination was investigated preincubating 50% platelet poor normal plasma at 37 o C for 30 min with Mab material. To 400 ~1 of this mixture was added 100 ~1 of platelet-rich plasma (adjusted to approx. 300 X 109/1), and agglutination recording was initiated after addition of ristocetin (Lundbeck, Copenhagen, Denmark) (final reaction concentration 1.5 mg/mI) in a Payton dual-chapel a~egometer.

69

Testing of Mabs for reactivity against VJII : C and VIII : Ag

Mabs were tested for VIII : C inhibiting activity following the Bethesda recommendations [35]. Mabs were further tested for affinity towards VIII : Ag in an ELISA in which VIII : Ag had been extracted from normal EDTA-plasma by a solid-phase immobilized human haemophilia A VIII : C neutralizing antibody [36]. In competitive sequential ELISA, serial dilutions of Mabs were introduced in a step following extraction of VIII: Ag from EDTA-plasma and free VIII: Ag epitopes were estimated using a HRP-conjugate of F(ab’)~-fra~ent of a haemop~~c antibody. Fast protein liquid chromatography (FPLC)

Citrated (1 part of trisodium citrate 3.8%, 9 parts of blood) normal plasma was diluted 1: 2.5 with PBS adjusted to pH 6.8. After sterile filtration (Millipore GV, 0.22 pm), 500 ~1 of this mixture was applied on a Mono-Q column HR 5/5 (Pharmacia, Uppsala, Sweden). High pressure liquid ~~omato~aphy was performed using an electronically controlled dual pump system (P-500, Pharmacia) and 214 nm light absorbancy recording. Buffers (0.02 mol/l Tris, pH 7.2: buffer A; 0.02 mol/l Tris, pH 7.5, with 1 mol/l of NaCl: buffer B) were mixed through pump delivery to form a linear gradient of sodium chloride ranging from O-l mol/l. The flow was 2 ml/min. Fractions (480 ~1) were collected at 0.2~min intervals, and loo-p1 portions of each fraction was pipetted into wells of microtitre plates precoated with rabbit-~ti-vWf : Ag. Bound material was detected by HRP-conjugated Mabs or HRP-conjugate of the polyclon~ antibody. Substituting normal plasma, titrated plasma from a patient with hz-vWd was used in control experiments. Evaluation of binding and detection efficiencies of Mabs

All seven Mabs were tested for binding and detection efficiency by ELISA. Mab immunoglobulins (approximately 2 pg/ml) were used for coating of microtitre plates. Serial dilutions of a fixed panel of different patient material (hz-vWd, vWd type IIA, dominant vWd type I, heterozygous recessive vWd) was used, and retained material was detected by HRP-conjugates of each of the seven individual Mabs in experiments covering all combinations. Development of two-site Mab based ELISA for plasma v Wf: Ag

During initial screening it was found that Mab vWf-33 gave decreased extraction of patient material of vWd type IIA. Hence, this antibody was investigated in two-site Mab ELISA, using vWf-41 at the detection side, and compared with the opposite construction of the sandwich and with the polyclonal ELISA. We tested plasma samples from 20 healthy persons (10 males, 10 females), 13 patients with type I vWd and 11 patients with various type II vWd (some patient material of subtype II was kindly donated by Prof. I.M. Nilsson, Malmti and Prof. I. Scharrer, Frankfurt). The assay protocol was as before [31]. In brief, Mab vWf-33 diluted to 2-4 pg/ml in 0.1 mol/l of carbonate buffer (Na,CO, 3.18 g, NaHCO, 5.86 g/l), pH

70

9.6, was pipetted into all wells of the Nunc Immunoplate I (Nunc, Copenhagen, Denmark) microtitre plate and incubated at 4°C overnight. After three washes in ELISA washing buffer, duplicates of 100 ~1 of serially diluted normal plasma ranging from 1: 10 to 1: 2 560 in PBS-Tween supplemented with 50 g/l of bovine albumin (Sigma A-7030) were applied. In most instances, material from patients with vWd subtypes II were applied in a number (4-6) of different dilutions for study of dose-response congruity. The extraction phase took place at 4°C for 14-18 h. After three washings as above, 100 ~1 of appropriately diluted HRP-conjugate (Mab vWf-41, or polyclonal antibody) in PBS-Tween with 10 g/l of bovine albumin was pipetted into each well, and after incubating for 2 h at room temperature the plate was washed three times as before. One hundred microlitres of substrate mixture of o-phenylendiamine and urea-peroxide in 0.1 mol/l citrate-phosphate as above was added and, after 4-15 min the enzyme reaction was stopped by addition of 100 ~1 1 mol/l of sulphuric acid. The absorbance was read at 492 nm in the automated ELISA reader, Titertek Multiskan MCC/340 (Flow Laboratories, Irvine, Scotland, UK). Results Detection and extraction characteristics Immunoglobulin from eight hybridoma supematants bound to the immunogen. Seven Mabs proved specific on further testing. The immunoglobulin isotype of each Mab is listed in Table I. The selected hybridoma material showed high affinity towards normal plasma vWf (Fig. 1). In sequential competitive ELISA and at high concentrations, all Mab material exhibited some cross-reactivity with the polyclonal

TABLE

I

Characteristics Mab code

vWf-20 VWf-21 vWf-23 VWf-25 vWf-33 VWf-39 VWf-41 ’ b c *

of seven monoclonal

Immunoglobulin isotype

IgG, Igct I@, IgG,, IgG, IgG, IgG,

antibodies

Mab efficiency Binding

+++ ++++ ++ ++ ++++ +++ ++++

a

Detection

+++ ++++ +++ ++ +++ +++ ++++

(Mabs)

against

Inter-Mab competition

0 0 21 (weakly) 0 0 0 0

b

vWf (vWf : Ag) VIII : c mhibition (BU/mg)

0 0 1.3 0 0 0 1.3

’

Ristocetin cofactor inhibition

Collagen binding inhibition

(S) *

(S&j*

17 40 0 40 0 72 0

0 0 0 50 50 0 100

Binding and detection efficiency are expressed semi-quantitatively from data obtained by ELISA tests. Inter-Mab competition was recognized by cross-testing in sequential competitive ELISA. VIII : C inhibition is expressed in Bethesda U/mg immunoglobulin. Details of procedures for testing of inhibition of the ristocetin cofactor and of the vWf binding to collagen are explained in ‘Materials and Methods’.

71

A

492

2.0 -

1.0

-

5

3200

1600 reciprocal

600

400

of supernatant

200

100

dilution

Fig. 1. Primary screening of hybridoma supematants for antibody activity against von Willebrand factor antigen (vWf :Ag). Microtitre plates coated with rabbit-anti-vWf (10 pg/ml) extracted vWf :Ag from diluted normal plasma. Varying dilutions of hybridoma supematant material of clones; vWf-21, r; vWf-23, o; vWf-25, a; vWf-33, H; and vWf-41, 0 were reacted and subsequently, bound mouse immunoglobulin was detected by peroxidase conjugated anti-mouse immunoglobulin.

antibody (Fig. 2), presumably due to steric hindrance at some binding sites required by the polyclonal antibody. Patterns similar to those shown in Fig. 2 were found for all seven Mabs. Using hybridoma supematants for extraction of plasma vWf and F(ab’),-HRP conjugate of the polyclonal antibody for detection, good dose-response relationships were found (Fig. 3) with most of these antibodies. No binding to hz-vWd plasma was found in different two-site Mab ELISA constructions, some of which are presented in Fig. 4. Also, extremely low background readings were observed. The distinct specificities of these Mabs was demonstrated in competitive sequential assays showing a slight displacement by Mab vWf-21 at binding sites required by Mab vWf-23, whereas no other competitive interaction was identified in tests covering all combinations (Table I).

Analysis of FPLC fractions of normal plasma By anionic ion-exchange fast protein liquid chromatography, normal plasma vWf : Ag was separated from most of the UV-absorbing plasma proteins as shown in

72 A

492

40

lb

2’0

d

reciprocal of supematant dilution Fig. 2. Displacement of rabbit polyclonal-anti-vWf : Ag antibody by hybridoma supematants. Microtitre plates coated with rabbit anti-vWf immunoglobulin (10 pg/ml) were incubated with 1: 10 diluted normal plasma. Hybridoma culture supematant material at varying dilutions were incubated (vWf-21, O; vWf-23, A; and vWf-33, n) and peroxidase conjugated F(ab’), fragment of rabbit-anti-vWf : Ag detected free vWf : Ag epitopes. I), anti-TNP.

A 492

lb

do

160

640

2560

reciprocal of plasma dilution Fig. 3. Specificity testing of hybridoma supematant material in ELISA for vWf: Ag. Microtitre plates coated with rabbit anti-mouse ~uno~obulin were incubated with 1: 106 dilutions of hybridoma su~mat~ts of vWf-21, o; vWf-33, 8. Serial dilutions of normal plasma and hz-vWd plasma were reacted. Bound vWf : Ag was detected by peroxidase-labelled F(ab’), fragment of rabbit anti-vWf : Ag. A: hz-vWd plasma (vWf-41, other antibodies reacted sirnilariy).

73

reciprocal of plasma dilution Fig. 4. Two-site ELISA dose-response of plasma using ascites hybridoma immunoglobulin for extraction and peroxidase (HRP) labelled hybridoma immunoglobulin for detection. A, homozygous recessive von Willebrand’s disease plasma. o, vWf-33 for binding and vWf-21 HRP for detection; 0, vWf-33 for binding and vWf-41 for detection; q, vWf-21 for binding and vWf-41 for detection.

Fig. 5, demonstrating the elution profile with two-stage assay by the polyclonal antibody and, for comparison, the identical chromatographic elution pattern using HRP-conjugates of two purified Mabs. In no case did Mabs bind to fractions of hz-vWd plasma. Equal results were found with the other Mabs (data not shown).

Binding to v Wf multimers Nitrocellulose blots of SDS-agarose gel electrophoresed normal plasma and plasma from a patient with hz-vWd run in parallel were incubated with individual Mabs. All of these bound equally well to the full range of multimers of normal plasma, whereas no stain was found in lanes containing hz-vWd material. The multimeric pattern of normal plasma as detected with four different Mabs is shown in Fig. 6A. Other Mabs reacted similarly. Figure 6B is an enlargement of the oligomer of lowest molecular weight of different type II vWd material detected by (a) polyclonal antibody, and (b) Mab vWf-33, showing that the two antibodies in principle identify the same subband pattern.

74

A

492

NaCl mot/l

c

5

15

25

35

45

55

65

75

85

95

Fig. 5. FPLC profile of vWf : Ag of normal plasma. 500 ~1 of 1: 2.5 diluted normal plasma was applied to a Mono-Q HR 5/5 colunm. Elution was performed by a linear gradient from 0 to 1 mol/l of sodium chloride in 0.02 mol/l Tris, pH 7.5. Flow was 2 ml/mm. Fractions (480 ~1) were applied (100 ~1) to three microtitre plates coated with 10 ng/ml rabbit anti-vWf : Ag antibody. Following extraction, peroxidase labelled F(ab’), fragment of rabbit anti-vWf : Ag ( x x x x x X) or peroxidase labelled ascites hywere incubated. Absorbancy at 214 bridoma i~uno~ob~ of vWf-41 (.- .-.) and vWf-23 ( -) ). Ordinate : o.d. 492 recordings of fractions in ELISA for nm (- - - - - -). Sodium chlonde gradrent (vWf : Ag.

Mab interaction with the ristocetin induced platelet agglutination Four antibodies partially inhibited the platelet agglutination by inducer ristocetin (vWf-20, vWf-21, vWf-25, vWf-39), as listed in Table 1.

Fig. 6. A. Multimeric vWf pattern of normal plasma usmg four different monoclonal antibodies (Mabs). Normal plasma was separated by SDS-agarose gel and transferred onto nitrocellulose. Blots were incubated with Mabs (direct technique: horse-radish peroxidase conJugates vWf-33 and vWf-41; indirect technique: monoclonal immunoglobulin of vWf-23 and vWf-25 followed by peroxidase conmgate of rabbit anti-mouse immunoglobuhn). Oligomers were visualized by development with diaminobenzidineH,Oa. B. Nitrocellulose blot showing the lowest mol wt multimer of normal plasma (N) and plasma from patients with von Willebrand’s disease subtypes HA, IIB, IIC, and IIC after SDS-agarose gel electrophoreais followed by electrotransfer onto nitrocellulose. Blots were mcubated with: (a) horse radish peroxidase conjugate of rabbit anti-vWf (Dakopatts, code P-226; (b) Mab vWf-33, followed by biotinylated rabbn anti-mouse i~uno~obu~s, and subsequently peroxldase labelled avidm. Development was performed as in (A).

VWF-23

VWF-25

VWF-33

VWF-41

a.

NP

IIA

IIB

NP

IIA

IIB

IIC

IIC

IID

IID

TABLE II Plasma vWf : Ag (mean + SD) in: (A) 20 healthy blood donors (10 females, 10 males) and (B) 13 patients with type I von Willebrand’s disease PAb ELISA (U/ml) A

0.87 + 0.29 rd

B

’

vWf-33/vWf-41-HRP ELISA (U/ml) ’

0.87 f 0.26

0.97 f 0.26

0.952

0.44*0.11 r

vWf-4l/vWf-33-HRP ELISA (U/ml) b

r 0.43f0.11

0.911

r

0.192

0.47 5 0.14

0.847

a ELISA utilizing a polyclonal antibody, (PAb) and two monoclonal antibodies. b vWf-4l/vWf-33-HRP denotes vWf-41 for extraction and vWf-33 for detection. HRP = horse radish peroxidase coqugate. ’ vWf-33/vWf-41-HRP denotes MAb vWf-33 for extraction and vWf41 for detection. d r, coefficient of correlation.

Mab inhibition of v Wj binding to collagen vWf-41 seemingly completely blocked vWf adhesion to type I collagen, and two other Mabs showed some inhibitory activity against this binding function (vWf-25, vWf-33) (Table I). Cross-reactivity with other proteins ofplasma In no case did any Mab elicit binding to fibrinogen, fibronectin, qmacroglobulin or IgM in ELISA’s using monospecific polyclonal antibodies retaining and presenting these antigens from normal plasma.

Two-site Mab ELISA for plasma v Wj: Ag From experiments using various combinations of Mabs in two-site ELISA employing crude Mab immunoglobulin at the extraction side and HRP-conjugates of Mabs at the detection side, it was found that symmetrical ELISA gave very poor dose-dependent responses. However, in asymmetrical designs (using different Mabs at each side of the ELISA sandwich), steep dose-response relationships were encountered. Some of these are shown in Fig. 4. Further is was experienced, that Mab vWf-33 used for extraction gave decreased response to type IIA vWd material included in the screening procedure compared to polyclonal ELISA results. Table II shows results obtained with normal plasma samples and subtype I vWd material in two opposite assay constructions of Mabs vWf-33 and vWf-41 and polyclonal ELISA. Slightly higher values were recorded in normal plasma using vWf-33 for extraction and vWf-41 for detection. However, this was not seen with type I vWd material. Plasma from three severe homozygous (or doubly heterozygous) vWd patients gave values below 0.01 U/ml of vWf: Ag in all three assays. Typical standard curves of the two-site Mab ELISA’s described here are shown in Fig. 7. Further investigation of the assay model using vWf-33 for extraction gave consider-

77

I

I

10

I

1

160

40

I

2560

640

PLASMA DILUTION Fig. 7. Normal plasma standard curves for two-site monoclonal antibody ELISA’s for von WiIIebrand factor (vWf: Ag) in plasma. 0, monoclonal antibody (Mab) vWf-41 for extraction and peroxidase conjugate of Mab vWf-33 for detection; 0, Mab vWf-33 for extraction and peroxidase conjugate‘of vWf-41 for detection. TABLE

III

Plasma von Willebrand factor antigen (vWf: Ag) in patients with von Willebrand’s disease (vWd) subtypes II measured by ELISA using two-site polyclonal antibody ELISA (PAb) and monoclonal antibodies (vWf-33 and vWf-41) in different constructions Patient no.

vWd subtype

I II III IV V VI VII VIII IX X XI

IIA IIA IIA IIA IIA IIA IIB IIB IIB IIC IID

PAb ELISA

vWf-4l/vWf-33-HRP Mab ELISA (U/ml) ’

vWf-33/vWf-41-HRP Mab ELISA (U/mI) b

vWf-33/PAb-HRP ELISA (U/ml) ’

0.15 0.12 0.33 0.38 0.53 0.34 0.30 0.14 0.30 0.39 0.89

0.05 0.06 0.11 0.12 0.16 0.09 0.08 0.09 0.12 0.10 0.26

0.06 0.05 0.11 0.11 0.15 0.09 0.09

(U/ml) 0.19 0.16 0.33 0.34 0.53 0.34 0.25 0.16 0.33 0.41 1.25

0.09 0.12 d 0.22 d

a vWf-4l/vWf-33-HRP denotes: vWf-41 for extraction and vWf-33 for detection b vWf-33-vWf-41-HRP denotes: vWf-33 for extraction and vWf-41 for detection. ’ vWf-33/PAb denotes: vWf-33 for extraction and HRP-conjugate of polyclonal antibody (Dakopatts, code P 226) for detection. d Two samples reacted incongruent with the standard curve. Estimates were based on 1: 10 dilution only.

78

ably decreased values of vWf : Ag for all type II vWd material tested (Table III). Subsequent experiments using vWf-33 as solid phase antibody and the polyclonal antibody as detector, resulted in almost identical recordings. We only observed incongruity (previously denoted non-parallelism) in vWd subtypes IIC and IID using the HRP-conjugate of undigested IgG of the polyclonal antibody. In a previous study using F(ab’), fragment, these plasmas reacted congruently with normal plasma [31]. Discussion The present report describes the production and characterization of seven murine Mabs against human vWf. In functional assays we identified four Mabs (vWf-20, vWf-21, vWf-25 and vWf-39) interacting with ristocetin induced platelet agglutination, suggesting specificity for the site of the vWf polypeptide binding to platelet glycoprotein Ib [37]. One Mab (vWf-41) strongly inhibited vWf binding to collagen, vWf-33 partly inhibited this function of vWf, and vWf-25 showed some reactivity against both ristocetin agglutination and collagen binding of vWf. The antigenic sites of one previously described collagen binding site and the ristocetin binding site apparently are related to each other and located near the N-terminal part of the vWf subunit [37, and review 381. Recently a second binding site for collagen was discovered [15] in a site homologous with the first detected collagen binding epitope [14,37]. The epitope binding to platelet glycoprotein IIb/IIIa complex of the activated platelet is presumed to be located at the C-terminal part of the vWf subunit peptide [38]. vWf seemingly possess polyvalent binding characteristics, as other binding properties have been reported, such as FXIIIa dependent cross-linking to fibrin [39] and adhesion to non-collagenous subendothelial proteins [41]. The relative importance of these individual binding effects in formation of the haemostatic plug is still not clear. Using the detection pattern of the polyclonal antibody as reference, the present series of Mabs bound similarly to vWf of normal plasma fractionated by FPLC and to all multimers of normal plasma. The majority of Mabs had favourable extraction characteristics towards plasma vWf. The series of Mabs maintained their excellent detection properties after conjugation with peroxidase for use in multimeric sizing techniques and in ELISA for vWf : Ag. Based on results of competitive and non-competitive ELISA. it is assumed, that this series of Mabs covers a number of different vWf-epitopes. We constructed different two-site ELISA’s for plasma vWf : Ag, and found poor response when the same antibody was used in both stages of the assays. Asymmetrical applications of the antibodies showed high sensitivity compared with the only previously reported two-site Mab based ELISA [23], in which Mabs recognizing the glycoprotein Ib binding epitope of vWf were used. In this, Chand and coworkers showed that subtype IIA vWd material had decreased reactivity in two-site Mab ELISA compared with results obtained by rocket electrophoresis. In the present study, the antibody vWf-33, not inhibiting the ristocetin induced platelet agglutination but interfering with collagen binding of vWf, gave decreased extraction of vWf : Ag from subtype II vWd origin. This antibody detected all multimers of normal plasma and retained vWf from

19

normal plasma and from plasma from subtype I vWd patients equally well as the polyclonal antibody. Further, the vWf-33 seemingly detected the major bands and subbands of various type II vWd origin in a fashion similar to the polyclonal antibody. Thus, the particular low response to type II vWd material is not solely due to the defective polymerization that hallmarks vWf in subtypes II vWd. It is assumed, that other common defects might exist in vWf from such patients. Some indication exist pointing in the same direction [42]. However, further investigative work is needed to prove the validity of this hypothesis. Acknowledgements The technical assistance of T.S. Andreasen, G. Krog, B. 0rum, B. Carlsen, K. Christiansen, I.L. Hansen and A. Christensen is greatly appreciated. This work was supported by The Danish Blood Donors Foundation and The Institute of Experimental Clinical Research, University of Aarhus. References 1 Jaffe EA, Hoyer LW, Nachman RL. Synthesis of von Willebrand factor by cultured human endothelial cells. Proc Nat1 Acad Sci (USA) 1974;71:1906-1909. 2 Ginsburg D, Handin R, Bonthron DT, Donlon TA, Bruns GAP, Latt SA, Orkin SH. Human von Willebrand factor (vWf): isolation of complementary DNA (cDNA) clones and chromosomal localization. Science 1985;228:140-1406. 3 Vetweij CL, De Vries CJM, Distel B, et al. Construction of cDNA coding for human von Willebrand factor using antibody probes for colony-screening and mapping of the chromosomal gene. Nucl Acid Res 1985;13:4699-4717. 4 Meyer D, Obert B, Edgington T. Monoclonal antibodies specific for factor VIII from cellular hybrids. Circulation 1980;62 (Suppl III):106 (abs.). 5 Sola B, Avner P, Sultan Y, Jeanneau C, Maissonneuve P. Monoclonal antibodies against human factor VIII molecule neutralize antihemophihc factor and ristocetin cofactor activities. Proc Nat1 Acad Sci (USA) 1982;79:183-187. 6 Ogata K, Saito H, Ratnoff OD. The relationship of the properties of antihemophilic factor (Factor VIII) that support ristocetin-induced platelet agglutination (Factor VIIIR: RC) and platelet retention by glass beads as demonstrated by a monoclonal antibody. Blood 1983;61:27-35. 7 Meyer D, Zimmermann TS, Obert B, Edgington TS. Hybridoma antibodies to human von Willebrand factor. I. Characterization of seven clones. Br J Haematol 1984;57:597-608. 8 Stel HV, Sakariassen KS, Scholte BJ, et al. Characterization of 25 monoclonal antibodies to factor VIII-von Willebrand factor: Relationship between ristocetin-induced platelet aggregation and platelet adherence to subendothelium. Blood 1984;63:1408-1415. 9 Meyer D, Baumgartner HR, Edgington TS. Hybridoma antibodies to human von Willebrand factor. II. Relative role of intramolecular loci in mediation of platelet adhesion to the subendothelium. Br J Haematol 1984;57:609-620. 10 Sixma J, Sakariassen KS, Stel HV. et al. Functional domains on von Willebrand factor. J Clin Invest 1984;14:736-744. 11 Goodall AH, Jarvis J, Chand S, et al. An immunoradiometric assay for human factor VIII/van Willebrand factor (VIII : vWf) using a monoclonal antibody that defines a functional epitope. Br J Haematol 1985;59:565-577. 12 Ruggeri ZM, De Marco L, Gatti L, Bader R, Montgomery RR. Platelets have more than one binding site for von Willebrand factor. J Clin Invest 1983;72:1-12.

80 13 De Marco L,, Girolami A, Zimmermann TS, Ruggeri ZM. Von Willebrand factor interaction with the glycoprotein IIb/IIIa complex. J Clin Invest 1986;77:1272-1277. 14 Girma J-P, Kalafatis M, Pietu G, et al. Mapping of distinct von Willebrand factor domams interacting with platelet GPIb and GPIIb/IIIa and with collagen using monoclonal antibodies. Blood 1986;67:1356-1366. 15 Roth GJ, Titani K, Hoyer LW, Hickey M. Localization of binding sites within human von Willebrand factor for monomeric type III collagen. Biochemistry 1986;25:8357-8361. 16 Ftqimura Y, Titani K, Holland LZ, et al. A heparin binding domain of human von Willebrand factor. J Biol Chem 1987;262:1734-1739. 17 Foster PA, Fulcher CA, Marti T, Titam K, Ztmmermann TS. A maJor factor VIII bmding domain resides withm the amino-terminal 272 ammo acid residues of von Willebrand factor. J Biol Chem 1987;262:8443-8446. 18 Sultan Y, Avner P, Maissonneuve P, Amaud D, Jeanneau Ch. An unmunoradiometric assay for factor VIII related antigen (VIIIRAg) using two monoclonal antibodies - comparison with polyclonal rabbit antibodies for use in von Willebrand’s disease dtagnosis. Thromb Haemostas 1984;52:250-252. 19 Larmne S, Wallmark A, Holmberg L, Nilsson IM, Sjogren H-O. The use of monoclonal antibodies in measuring factor VIII/van Willebrand factor. Stand J Clin Lab Invest 1985;45:17-26. 20 Thomas JE, Peake IR, Giddmg JC, Welch AN, Bloom AL. The application of a monoclonal antibody to factor VIII related antigen (VIIIRAg) in immunoradiometnc assays for factor VIII. Thromb Haemostas 1985;53:143-147. 21 Bradley LA, Franc0 EL, Reisner HM. Use of monoclonal antibodies m an enzyme immunoassay for factor VIII-related antigen. Clin Chem 1984;30:87-92. 22 Inoue K, Mazda T, Tobe T, Tomita M. Enzyme-linked immunosorbent assay for von Willebrand factor antigen using a monoclonal antibody. Chem Pharm Bull 1986;2550-2555. 23 Chand S, McCrew A, Hutton R, Tuddenham EGD, Goodall AH. A two-site, monoclonal antibodybased immunoassay for von Willebrand factor - demonstration that vWf function resides m a conformational epitope. Thromb Haemostas 1986,55:318-324. 24 Lynch D, Zimmerman TS, Collins CJ, et al. Molecular cloning of cDNA for human von Willebrand factor: authentication by a new method. Cell 1985;41:49-56. 25 Van der Kwast TH, Stel HV, Cristen E, Bertma RM, Veerman ECI. Localization of factor VIII-procoagulant antigen: an immunohistological survey of the human body using monoclonal antibodies. Blood 1986;67:222-227. 26 Goodall AH, Meyer D. Registry of monoclonal antibodies to factor VIII and von Willebrand factor. Thromb Haemostas 1985;54:878-891. 27 Keamey JF, Radbruch A, Liesegang B, Rajewsky K. A new mouse myeloma cell lure which has lost imrnunoglobuhn expression but permits the construction of antibodysecreting hybnd cell lines. J Immunol 1980;123:1548-1555. 28 Kiihler G, Milstein C. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature (London) 1975;256:495-497. 29 Kiihler G, Milstein C. Derivation of specific antibody-producing tissue cultures and tumor cell lines by cell fusion. Europ J Immunol 1976;6:511-519. 30 Nakane PK, Kawaoi A. Peroxidase-labelled antibody. A new method of coqugation. J Histochem Cytohistochem 1974;22:1084-1091. 31 Ingerslev J. A sensitive ELISA for von Willebrand factor (vWf: Ag). Stand J Clin Lab Invest 1987;47:143-149. 32 Bukh A, Ingerslev J, StenbJerg S, Moller NPH. The multtmeric structure of plasma FVIII: RAg studied by electroelution and immunoperoxidase detection. Thromb Res 1986;43:570-584. 33 Brown JE, Boska JO. An ELISA test for the binding of von Willebrand factor anttgen to collagen. Thromb Res 1986;43:303-311. 34 Shulman M, Wilde CD, Kiihler G. A better cell line for making hybridomas secreting specific antibodies. Nature 1978;276:269-270. 35 Kasper CK, Aledort LM, Counts RB, et al. A more uniform measurement of factor VIII inhibitors. Thromb Diath Haemorrh 1975:34:869-872.

81 36 Ingerslev J, Stenbjerg S. Enzyme linked immunosorbent assay (ELISA) for the measurement of factor VIII coagulant antigen (CAg) using haemophilic antibodies. Br J Haematol 1986;62:325-332. 37 Girma J-P, Chopek MW, Titani K, Davie EW. Limited proteolysis of human von Willebrand factor by Staphylococcus aurtw V-8 protease; isolation and partial characterization of a platelet binding domain. Biochemistry 1986;25:3156-3163. 38 Shelton-Inloes BB. Titani K, Sadler JE. cDNA sequences for human von Wtllebrand factor reveal five types of repeated domains and five possible protein sequence polymorphism. Biochemistry 1986;25:3164-3171. 39 Hada M, Kaminski M, Bockenstedt P, McDonagh J. Covalent crosslinking of von Willebrand factor to fibrin. Blood 1986;68:95-101. 40 Pareti F, Fukimura Y, Dent JA, Holland LZ, Zimmermann TS, Ruggeri ZM. Isolation and characterization of a collagen bindmg domain in human von Willebrand factor. J Biol Chem 1986;261:15310-15315. 41 Wagner DD, Urban-Pickering M, Marder VJ. Von Willebrand protein binds to extracellular matrices independently of collagen. Proc Nat1 Acad Sci USA 1984;81:471-475. 42 Takahashi Y, Kajafatis M, Girma JP. Meyer D. Abnormality of the N-terminal portion of von Willebrand factor (Fragment Sp III) in type HA and IIC von Willebrand’s disease. Thromb Haemostas 1987;58:499 (abs.).