Haemocompatible surfaces Field of the Invention The present invention in general relates to the involvement of non- adhesive proteins such as human serum albumin (HSA) in the prevention of cell attachment to non-biological surfaces such as polymeric plastics, metals, ceramics etc. More particularly, the present invention relates to monoclonal antibodies directed against HSA and to the use of such monoclonal antibodies in the production of a biomaterial. The present invention also relates to a method for the coupling of antibody to a solid support to minimise non-specific adsorption of antibody to the surface.

Background of the Invention One of the major difficulties presently encountered in the preparation of biomaterials for certain applications is that of preventing cellular attachment to the biomaterial. It is believed that the design of such biomaterial surfaces would be advantaged if non-adhesive proteins could be specifically bound and held with higher affinity by the biomaterial surface.

It would also be of advantage if the proteins adsorbed onto the biomaterial surface were derived from the plasma or serum of the patient, and that the plasma or serum could be used without the requirement to initially purify the non-adhesive proteins from the plasma or serum.

The biological activity of a surface that has adsorbed a non-adhesive protein is dependent upon the way in which the domain or domains of the protein are presented. It is believed that the ability to modify the biological activity of a surface to which non-adhesive proteins are adsorbed by choosing the orientation in which they are bound would also be advantageous.

Furthermore, the adsorption of a protein to a surface may alter its conformation, thereby altering its biological activity 1 and forcing for example a normally non-adhesive protein such as HSA to adopt a conformation which promotes cell attachment The use of a monoclonal antibody to enable specific orientated binding overcomes this problem.

Summary of the Invention The present inventors have produced a hybridoma cell line which produces monoclonal antibodies reactive with HSA.

Accordingly, in a first aspect the present invention consists in a monoclonal antibody reactive with HSA in a manner such that when the

monoclonal antibody is bound to HSA the HSA retains its ability to prevent cell attachment.

The present invention further consists in a hybridoma cell line producing such a monoclonal antibody.

This monoclonal antibody has been designated 20/13 and the hybridoma cell line producing monoclonal antibody 20/13 has been designated PE20/13A11. This hybridoma cell line has been deposited with the European Collection of Animal Cell Cultures (ECACC), Vaccine Research and Production Laboratory, Public Health Laboratory Service, Centre for Applied Microbiology and Research, Porton Down, Salisbury, Wiltshire SP4 OJG, United Kingdom, on 22nd May 1996 and accorded accession number PE20/13A11 96052171.

In a second aspect the present invention consists in a biomaterial including a support, the biomaterial being characterised in that at least one surface of the support is coated with monoclonal antibody 20/13.

In a preferred embodiment of this aspect of the present invention the support is a polymer selected from the group consisting of agarose, polyacrylamide, polyvinyl alcohol, polyester, polystyrene, polytetrafluoroethylene, expanded polytetrafluoroethylene, poly (ethylene terephthalate), polyurethane, silicon rubber and poly (hydroxyethyl methacrylate).

In a further preferred embodiment of this aspect of the present invention the support is coated with metal, lipid or defined chemical moieties.

In a further preferred embodiment of the present invention the support is a preformed membrane, a porous material, a sponge or a tube.

By appropriate use of the monoclonal antibody 20/13 described herein, it is possible to produce a biomaterial, the surface of which will specifically bind HSA in a manner such as to repel the attachment and growth of specific cells.

In the production of the biomaterial of the second aspect of the present invention, the monoclonal antibody may be linked to any surface using techniques known to a person skilled in the art. This technique will obviously vary from material to material as would be readily understood, however, the general concept of the present invention enables the use of any

material to which the monoclonal antibodies may be linked to produce a biomaterial.

The attachment of the monoclonal antibody to the support may be achieved by covalent coupling, radiation grafting or adhesive bonding. In addition, the attachment of the monoclonal antibodies to the support may be achieved by use of affinity binding to a suitable affinity matrix, such as anti- mouse Ig matrix or Protein A.

In a preferred embodiment of this aspect of the present invention the coupling methodology uses a robust two-component approach such as the avidin-biotin system. In this embodiment one component (which may be avidin) is covalently coupled to the solid support, then non-coupled material stripped from the surface and the surface rendered incapable of mediating further protein adsorption. The non-coupled material may be stripped by rinsing with a denaturing detergent. The denaturing detergent may include, for example, 0.1-10% sodium dodecyl sulphate (SDS) and. optionally mercaptoethanol. The second component (which may be biotin) is covalently coupled to the monoclonal antibody, and any non-coupled material removed. When the derivitised surface and antibody are brought together under conditions as to promote the interaction of the coupling components, the antibody becomes specifically coupled to the surface. In this way there is no opportunity for the adventitious (non-covalent) adsorption of antibody to the surface. This is of significance in biomaterials applications, where leaching of a biological moiety may have deleterious effects.

In addition, as would be readily understood, the species of animal from which the monoclonal antibody is produced is preferably compatible to that in which the biomaterial is to be implanted, ie. if the material is to be implanted in humans, humanised monoclonal antibodies or fragments thereof having antigen-binding activity are preferably used.

The observations for the role of non-adhesive proteins in prevention of cellular attachment to polymeric plastics are of direct relevance to the design of biomaterials, and point to another application of the invention.

The application is to improve that biological compatibility of biomaterial surfaces or to minimise cell colonisation of biomaterial surfaces. For this purpose, the invention describes the use of monoclonal antibodies, either those described in the Examples of this invention or others, in immobilised

form on the surface, so that the antibody will bind non-adhesive proteins in a specific manner and with high affinity, so presenting on the surface a biological signal that will be recognised by other biological molecules and/or repel the attachment and growth of specific cells. The ability by this method to control the nature of cellular attachment to the surface of a biomaterial will overcome or alleviate some of the disadvantages of the existing art.

The design of the present invention has the additional advantage that the adsorbed proteins could be derived from the plasma or serum of the patient, and the plasma or serum could be used without requirement to purify the non-adhesive components from the plasma or serum. In other words, the introduction of a biomaterial of the present invention into the body of a host may result in the control of attachment mediated via adsorption of the host's own protein.

Brief description of the Figures Figure 1. Binding of anti-HSA antibodies to HSA Subscript: Binding of Mab 20/13 to purified HSA antigen coated at 0.005- 10.0, ug/ml.

Figure Z. Binding of HSA from plasma to immobilised anti-HSA antibodies Subscript: Binding of HSA from human plasma to purified anti-HSA Mabs coated at dilutions from 10.0 to O. Ogg/ml.

Figure 3. Binding of HSA from plasma to immobilised biotinylated anti- HSA antibodies Subscript: Binding of HSA from human plasma to purified biotinylated anti- HSA Mabs coated at dilutions from 10.0 to O. 0, ug/ml.

Figure 4. Effect of stripping procedure on surface density of immobilised avidin.

Subscript: Amount of avidin remaining on glutaraldehyde-activated surfaces following treatment with two regimes designed to remove unattached protein.

Figure 5. Effect of SDS on the biotin-binding capacity of immobilised avidin.

Subscript: Binding of biotinylated horseradish peroxidase to avidin immobilised on various substrates and treated with 4% SDS to remove unattached protein.

Figure 6. Antigen-binding activity of antibody'constructs'immobilised on nitrocellulose membranes.

Subscript: Binding of HSA and fibronectin to biotinylated Mabs coupled to avidin immobilised on nitrocellulose membranes.

Figure 7. Platelet attachment to immobilised monoclonal antibodies in vitro Subscript: Binding of platelets from platelet rich plasma to biotinylated Mabs coupled to avidin immobilised on FEP Figure 8. In vitro attachment of radiolabelled platelets to antibody- modified polyurethane tubing Subscript : Amount of Cr51 labelled platelets adsorbing to tubing surface after 30min exposure to flowing whole human blood at a shear rate of 227.7S- 1.

Figure 9: Cell attachment to TCPS bearing immobilised monoclonal antibodies Subscript: Initial (90min) attachment of BHK cells to biotinylated Mabs coupled to avidin immobilised on TCPS Figure 10: Cell attachment to glass bearing immobilised monoclonal antibodies Subscript: Initial (90min) attachment of BHK cells to biotinylated Mabs coupled to avidin immobilised on glass coverslips.

Figure 11 : Fibrinogen adsorption to TCPS bearing immobilised monoclonal antibodies Subscript: Binding of biotinylated human fibrinogen from dilutions of human plasma to purified Mabs coated at 10. 0, ug/ml on TCPS.

Detailed Description of the Invention In order that the nature of the present invention may be more clearly understood preferred forms thereof will now be described with reference to the following non-limiting examples.

EXAMPLES Materials and Methods: Monoclonal antibodies to HSA Female BALB/c mice were immunised intraperitoneally with 0.33ml platelet extract prepared as described previously4 mixed with 0.33ml Freunds Complete Adjuvant per mouse. After 3-6 months animals received

similar booster injections without adjuvant. Those producing antibody reactive in enzyme-linked immunoassay (ELISAs) with platelet extract were rested for 6 weeks, given a further boost and used for cell fusion 3 days later.

Spleen cells were fused with Sp2/0 myeloma cells, and hybrid cultures were grown as previously described 5. Supernatants from confluent wells were screened against HSA by ELISA as previously described 6 with the following modifications. Incubations were done in 50gel per well. All washes were reduced from 5 min to 3 min. After incubation with peroxidase-conjugated rabbit anti-mouse immunoglobulins (RAM-HRP, Dako P0161) plates were washed four times with PBS before addition of 100y1 per well O-phenylene diamine substrate (OPD, Sigma). Colour development was terminated after 3-30 min by addition of 25gel per well of 0.4% sodium azide using a dropper.

Absorbances were read on a Dynatech MR600 plate reader at 450nm (ref. 630 nm). Positive hybridomas were further screened for activities against HSA by ELISA and Western blotting against HSA or dilutions of serum. Hybridomas were cloned by microscopic picking of single cells 7. Antibody subclasses were determined by immunodouble diffusion against subclass specific antisera (Litton Bionetics) or by ELISA on HSA using subclass specific peroxidase-conjugated second antibodies (Litton Bionetics). Large quantities of antibody were prepared by hybridomas as ascites in BALB/c mice primed 7-14 days earlier with 0.5 ml pristane (Aldrich). Antibody was purified from ascites fluid by affinity chromatography on protein A-Sepharose (Pharmacia) as previously described. Non-specific mouse antibody of various subclasses was prepared by fractionation of normal mouse serum on protein A Sepharose or from pools of anti-influenza monoclonal antibodies.

Binding of anti-HSA antibodies to HSA 96 well PVC microtitre plates (Dynatech laboratories Cat # 001- 0102101) were soaked with 200, u1/well of Phosphate Buffered Saline (PBS) pH 7.4 for 1 hour at room temperature. Wells were coated in triplicate with 50, 1 of HSA at l0Ilg/ml and doubling dilutions thereof in PBS. Background control wells were coated with PBS only. Plates were sealed with plastic film then incubated for 48 hours at 4°C. The HSA solution was aspirated and the wells rinsed once with PBS prior to blocking with 200well of 0.05% Tween 0 20 (Sigma Cat # P1379) in PBS (PBS/Tween) for 1 hour at 37 C. After aspirating the blocking solution 50bel of anti-HSA Mabs at 5, ug/ml in 0 PBS/Tween were placed in each well for 1 hour at 37 C. After aspirating the

antibody plates were rinsed three times with 200, u1/well PBS/Tween prior to adding 50, u1 per well of RAM-HRP 1 : 1000 in PBS/Tween and incubating at 0 37 C for 1 hour. The conjugate was aspirated and the plates washed with 200, 1/well PBS followed by l00, ul/well substrate solution (1. lmg/ml 2,2'- Amino-bis (3-ethylbenz-thiazoline-6-sulfonic acid; ABTS, Sigma). After 15mins incubation at room temperature colour development was measured on a microplate reader (Bio-Rad 3550) at a wavelength of 405nm against a 490nm reference.

Binding of Plasma HSA to immobilised anti-HSA Mabs ELISAs were carried out as above with the following modifications.

Plates were coated with lO » g/ml anti-HSA Mabs and doubling dilutions thereof. After blocking and rinsing 50AI of human plasma (from citrated whole human blood) was placed in each well for 1 hour at 37 C. After rinsing in PBS/Tween 50, u1 of Rabbit anti-Human Albumin polyclonal antibody (Dako, Cat No A001) diluted 1: 1000 in PBS/Tween were added to 0 each well and incubated for 1 hour at 37 C, followed by 50y1 per well of peroxidase conjugated Swine anti-Rabbit antibody (SwAR-HRP; Dako P0217) o 1: 1000 in PBS/Tween at 37 C for 1 hour. All biotinylated proteins used in this study were labelled with NHS-LC-Biotin (Pierce, 21335) for 2hrs on ice according to the manufacturer's instructions.

Covalent Attachment of Avidin to Plasma-Modified FEP Hydrazide groups were introduced into the surface of fluorinated ethylene-propylene (FEP) tape 12.7mm wide by 25, um thick (Du Pont ; FEP 100 Type A) by plasma discharge, as described previously 9. These strips were spiralled around the inner surface of glass bottles and roller mixed with 25% glutaraldehyde (Ajax Chemicals) for 4hrs at room temperature. After removal of the glutaraldehyde and rinsing with PBS the strips were roller mixed with 25jug/ml avidin in 100mM Na2C03 pH 9. 0 overnight at room temperature. Strips were washed with Na2C03 buffer and then incubated for 15min at room temperature with 100mM Na2C03 buffer containing 0.2mg/ml sodium borohydride to reduce unreacted aldehyde. After washing with Na2C03 buffer followed by PBS, non-covalently bound avidin was removed by incubation with 4% SDS for 15min at room temperature followed by extensive rinsing with PBS. The strips were then placed in a 24 well jig, such that the bottom of each 8.5mm diameter well comprised the FEP substrate. 4 wells were used for each treatment. To examine the

functionality of the coupled avidin, wells were blocked with 200AI PBS/Tween for 2hr at room temperature, then rinsed and incubated with biotinylated horseradish peroxidase (Pierce 2193X) diluted 1: 1000 in PBS/Tween for 1hr at room temperature. The activity of surface-bound peroxidase was determined colourimetrically using ABTS as described above.

Where measurement of the surface density of avidin was required, the protein was first radiolabelled with 1125 using the chloramine-T method 10.

Where platelet or cell attachment studies were to be performed, all reagents were sterilised by 0.22ym filtration and all procedures carried out in a laminar flow hood.

Attachment of Antibodies to Nitrocellulose Membranes HSA was conjugated to HRP by the periodate method 11 16mm diameter discs of Nitrocellulose paper (Schneider and Schull) were cut with a cork borer and one disc was placed into each well of a 24 well tissue culture dish. 300AI of a solution containing 25, ug/ml avidin in PBS were added to each well and incubated overnight at room temperature then aspirated and blocked with PBS/Tween for 2hrs at room temperature. After <BR> <BR> <BR> aspirating the Tween, a stainless steel fence 12 with an internal diameter of 6.7mm was placed into each well. 100tel NHS-biotinylated antibody were placed in the central well of the fence, and 400/-tl of a second biotinylated antibody introduced into the outer ring surrounding the central well. The antibodies (5, ug/ml in PBS/Tween) were coated for 2hrs at room temperature then aspirated, washed in block and split into two groups. The first group was incubated with 300, u1/well bovine fibronectin at 20, ug/ml in block for 90mins. After aspiration and washing, 300bai rabbit anti-fibronectin polyclonal antibody (Dako A245) diluted 1: 1000 in PBS/Tween was added to each well and incubated for 1 hr, then aspirated, washed and incubated with 300 1 per well SwAR-HRP 1: 1000 in PBS/Tween for 1 hr. The second group were incubated with 300, u1/well HSA-HRP conjugate diluted 1: 500 in block for 90min at room temperature. The conjugates were then aspirated from both groups, followed by 5 PBS washes. Enzyme activity was visualised using a chloronaphthol substrate and the discs photographed.

Platelet Attachment to Antibody-Modified FEP Surfaces 12.7mm wide FEP strips were coated with 5yg/ml Mab 23 (anti- Fibronectin), or Mab 20/13 at 1001/well, and 5jug/ml Mab overnight at room temperature. After aspirating and rinsing with PBS, the strips were clamped

in a 24 well jig, such that the bottom of each 8.5mm diameter well comprised the FEP substrate. 8 wells were used for each treatment. 100yl human platelet-rich-plasma was added to each well and incubated at 37°C for 10 min. Unbound platelets were removed by gentle washing and adherent cells fixed with 4% formalin (v/v) in PBS for 10 minutes at room temperature.

After aspirating the fixative and washing in PBS, cells were stained using 0.1% Eosin for 1hr in the dark at room temperature and visualised under the confocal microscope. 10 fields per well (ie 80 fields per treatment) were captured to disk and analysed on a Q570 image analyser for the number and area of adherent platelets per field.

In Vitro Platelet Attachment to Antibody-Modified Polyurethane Surfaces Radiolabelled platelets were prepared as previously described4').

Briefly, 450ml citrated whole human blood was obtained. Platelets were <BR> isolated, washed and radiolabelled with 51 Cr. They were then washed and reconstituted with platelet poor plasma (PPP) and red cells, to give a platelet count of approximately 2.5x105 platelets per ml of reconstituted blood. The reconstituted blood has a 50/50 erythrocyte/PPP ratio. The reconstituted <BR> <BR> <BR> blood containing 51 Cr-labelled platelets was termed'radiolabelled blood'.

55cm lengths of polyurethane tubing (Carbothane PC3595A) with an internal diameter of 1.727mm were coated with PBS or Mab 20/13 at 5, g/ml in PBS by drawing the solution into the tubes at 3ml/min and then recirculating at 0.5ml/min for at least 1hr at room temperature using a peristaltic pump. One end of each piece of tubing was connected to a 50ml syringe (Terumo SSTW087) and the other end connected to a similar syringe containing 36ml radiolabelled blood. This syringe was placed in a purpose-built 6-channel syringe pump capable of delivering flow rates of 4 to 30 ml/min, corresponding to wall shear rates of 228S to 1708S and maintained at 37°C. Nylon tubing (800/200/275/100) was used as a control substrate. For platelet studies, blood was passed through the tubing at 4 ml/min for 30min.

After this time the blood was removed, the tubing rinsed with PBS. The tubing was then folded to fit a gamma counting tube, and counted. The amount of radioactivity associated with the tubes was quantified by gamma counting the entire tube (Packard Cobra II Auto-gamma, model number D5003).

Cell Attachment to Antibody-Modified TCPS Surfaces The wells of a 24-well tissue culture dish were coated with 25yg/ml sterile avidin (300, u1/well) overnight at room temperature then aspirated and blocked with 0.05% sterile Tween-20 for 2hrs at room temperature.

Biotinylated antibodies were then coated onto the immobilised avidin using stainless steel fences, as described above for nitrocellulose. After 7 washes with PBS, each well was incubated with 3001LI DMEM containing 20% FBS and 20, g/ml bovine fibronectin for 1hr at 37°C. 300tel of a suspension containing 3 x 105 BHK cells in DMEM were then added to each well, swirled to mix, and incubated for a further 90min at 37°C. After aspirating unattached cells, the plates were washed with PBS and fixed with formol saline for 30mins at room temperature prior to staining with Methylene Blue (1% in 10mM borate buffer pH8.5) for 30mins. Free dye was washed off using borate buffer and the stained cells photographed. For microscopy, cells were stained with 0.1% Eosin as above and visualised under the confocal microscope. A parallel set of wells contained 16mm glass cover slips to examine cell attachment to a modified glass surface.

Fibrinogen Adsorption from plasma onto Antibody-Modified Surfaces The experiments were done on 96 well TCPS plates (Nunclon, Denmark; cat. no. 1 67008). Plasma was prepared from citrated whole blood collected into 4ml Vacuette tubes (Interpath, cat no 454075). Biotinylated fibrinogen was prepared using NHS-LC-Biotin (Pierce 21335) for 2hrs on ice as per the manufacturer's instructions and dialysed against 500 ml TRIS buffered saline containing lOmM ethylenediamine tetra-acetic acid (TBS/EDTA). Benzamidine was added to a final concentration of lOmM and protein concentration determined by BCA assay. For the adsorption studies, three rows of 96 well ELISA plates were coated overnight with loogl/well of 10, ug/ml of either anti-HSA (20/13) or anti-EGF (C11) antibodies. The anti- EGF Mab was selected for its lack of cross-reactivity with human fibrinogen.

Another three wells were left in PBS. The wells were then washed with PBS and 50AI of degassed buffer added. A solution containing 20% plasma in PBS and biotinylated fibrinogen to 10% of the plasma fibrinogen level was serially diluted 1 in 2 in degassed PBS. 100, ut aliquots were transferred to the 96 well plate in triplicate immediately after aspirating the buffer. Blank wells contained PBS alone. After incubating with the surface for 30min at room temperature, the wells were washed with 0.1% bovine serum albumin (BSA)

block in PBS, 4 times at 30sec intervals. 50y1 streptavidin-peroxidase (Amersham RPN 1231) diluted 1: 1000 in block were added to each well and left for at least 30min and at most 2h. The wells were washed with PBS and loo ttl ABTS was added to each well. The colour change of the substrate was then measured using an ELISA plate reader.

Results Binding of Mab 20/13 to HSA The binding of Mab 20/13 to HSA coated onto PVC is shown in Figure 1. Antibody binding showed the typical sigmoid curve, saturating above 0.6, ug/ml HSA and with a zone of independent binding between 0.02 and 0.16jug/ml HSA.

Binding of HSA to immobilised anti-HSA antibodies The ability of two immobilised monoclonal antibodies to bind HSA from plasma is shown in Figure 2. Binding by both antibodies was similar, with both showing a region of independent binding from 0.3 to 2.5yg/ml Mab. However, the curve was steeper for Mab 24/16 and the plateau higher, indicating a higher avidity for HSA.

When the antibodies were modified by the insertion of biotin molecules into available lysine residues on the antibody molecule, the situation was reversed, as shown in Figure 3. In this case the region of independent binding was from 0.3 to 1.25yg/ml for Mab 20/13, but was displaced to the right for Mab 24/14, giving a range of 0.6 to 2.5, ug/ml. This shift did not appear to result from decreased adsorption of Mab 24/16 due to insertion of biotin residues, since there was no difference between the two Mabs when probed with streptavidin-peroxidase (data not shown). Rather, the displacement indicates a decrease in the avidity of the antibody-antigen binding 14 and demonstrates that antibodies vary in their susceptibility to modification of functionally important lysine groups'. This is an important consideration in selecting an antibody suitable for covalent attachment via lysine residues, which are generally distributed throughout <BR> <BR> <BR> the antibody molecule 17. Molar ratios of NHS-biotin to protein from 2.2: 1 to 52: 1 have been used to biotinylate Mabs 18. A molar ratio of 22: 1 inserted 8- 14 moles NHS-LC-biotin per mole of Mab, and at this level all the biotinylated Mab bound to immobilised streptavidin. Binding of the biotinylated molecules decreased when higher biotin : protein ratios were

used 18. It is clear, then, that the biotinylation strategies have to be optimised for individual antibodies.

Covalent Attachment of Avidin to Plasma-Modified FEP The surface density of avidin bound to plasma-modified FEP was typically between 700 and 800 ng/cm2, well in excess of the theoretical level of 560 ng/cm2 required for monolayer coverage, assuming a molecular size of 5.6 x 5.0 x 4. 0 nm. This was presumably due to multilayering on the surface, potentially enhanced by avidin's high carbohydrate content which is known to increase its non specific adsorption. The concentrations of avidin and glutaraldehyde had been determined by titration to give maximal levels of adsorption (data not shown), and the concentration of avidin was well in excess of that required for surface saturation, enhancing the possibility of'piggybacking'phenomena by simple mass action. Treatment with 4% Sodium Dodecyl Sulphate (SDS) and 5% (3-mercaptoethanol (BME) has been shown to be effective in stripping a variety of adsorbed plasma proteins from a solid substrate and as shown in Figure 4, significantly reduced the amount of avidin present on the modified FEP substrate. Given that this procedure was designed to be stringent enough to remove near- irreversibly adsorbed proteins it is reasonable to conclude that the remaining avidin was effectively covalently attached to the surface. Since each of avidin's four subunits contains a disulphide bond which help maintain its structure 23 it would be expected that treatment with a powerful reducing agent such as BME will result in a loss of biotin-binding activity due to denaturation of the protein. A loss of biotin-binding activity is clearly unacceptable in the present application, so less stringent procedures were adopted. 4% SDS proved to be as effective as the SDS/BME reagent in removing non-coupled protein (Figure 4) and furthermore did not prevent the subsequent binding of biotin to the immobilised avidin (Figure 5). It is important to note that even in concentrations as low as 0.2%, denaturing detergents such as SDS will reduce the antigen-antibody binding reaction by over 90% 24 This precludes its use where Mabs are to be directly attached to a substrate, and points to a significant advantage of the present invention: by employing an indirect coupling procedure, leaching of surface-bound components can be significantly reduced by stringent stripping procedures which will nevertheless not result in a loss of activity of the coupled antibody.

Attachment of Antibodies to Nitrocellulose Membranes Having demonstrated that avidin can be coupled to a solid substrate, it is important to know that biotinylated antibodies still retain their antigen- binding activity when coupled to the immobilised protein. Figure 6 shows that antibodies immobilised to nitrocellulose in this way still retained their antigen-binding activity. Furthermore, by employing two different antibodies, the applicability of the methodology to a range of coupling situations is demonstrated. Antigen binding was demonstrated in different ways for the two antibodies. The ability of Mab 20/13 to bind albumin was determined directly, using an HSA-peroxidase conjugate. Binding of Fn to Mab 23 was determined indirectly, using a sandwich ELISA. Immobilised Fn also retained ability to bind gelatin25 (data not shown). This indicates that this site was not blocked when Fn bound to the immobilised Mab. This emphasises another feature of the present invention: binding of protein to a Mab-immobilised surface enables control over the orientation of the protein from a knowledge of the epitopes recognised by the Mab, and the protein maintains a non-denatured conformation (bound Fn retained its ability to bind gelatin). These results are consistent with those of other workers who have demonstrated that antibodies immobilised via the biotin-avidin system can retain greater functional activity than those adsorbed directly onto a surface, due to a reduction in surface-induced denaturation Platelet Attachment to Modified FEP Surfaces The conditions used in this assay were deliberately chosen to maximise the potential for attachment of platelets to the substrate : an enriched source of platelets was used (platelet-rich plasma), the incubation was static with the substrate at the bottom of the test well, and there was no pre-incubation with plasma or HSA to enable pre-adsorption of the protein prior to contact with platelets. Since it is known that platelets can attach to adsorbed fibronectin via their a5b1 integrin receptors 27 » 28 Mab 23 was used as a positive control. As shown in Figure 7, platelet attachment to FEP could be controlled by the nature of the immobilised Mab. As expected, the anti- Fn Mab increased both the number and area of the attached platelets, but the most significant feature was the ability of immobilised anti-HSA Mab to reduce both these parameters in comparison to the untreated control. Since spreading of adherent platelets correlates with platelet activation and thrombogenesis, it is significant that on the anti-HSA surface spreading was

minimal. At higher magnification adherent platelets were more rounded than on the other substrates with little sign of pseudopodia (data not shown) indicative of activation 29 30 Although it has been argued that a layer of attached but unactivated platelets may render a surface'passive'by presenting a native structure (the platelet's plasma membrane) to the bloodstream, it seems unlikely that the rounded platelets observed on the anti-HSA surface would perform this function in the present application.

Rather, it is more likely that this morphology is indicative of poorly attached platelets 32 In this assay, since both platelets and proteins (as plasma) were introduced to the substrate simultaneously, protein adsorption had to dominate cell attachment in order for platelet adherence to be controlled. <BR> <BR> <BR> <P>Although studies on the kinetics of protein adsorption 33 and cell attachment 34 indicate that this would be the case, the existence of competitive protein adsorption phenomena in plasma 35 and the ability of platelets to respond dynamically to a substrate 32, 36 mean that this is by no means self-evident.

However in the present study the nature of the protein layer was clearly able to influence concomitant platelet attachment. This is a significant finding, since in a preferred aspect of the present invention, the introduction of such a surface into the body results in the control of attachment mediated via adsorption of the host's own protein.

In Vitro Platelet Attachment to Antibody-Modified Polyurethane Surfaces The optimum conditions for immobilising antibodies on the luminal surfaces of polyurethane tubes were determined as above, using ELISA to quantify HSA binding activity (data not shown). The critical parameters of this part of the study were the use of whole blood and the existence of flow, which profoundly affects platelet attachment and activation. The shear rate of 228S-1 used in this study is equivalent to that found in veins, and <BR> <BR> <BR> represents a low rate 38. Platelet attachment is generally considered to increase with flow rate: at high shear rates this may be mediated by von Willebrand factor, and so shear conditions can be considered more rigorous than for static platelet adhesion tests. The radiolabelling method was chosen for its minimal impact on platelet function40. Although platelets can become activated during blood collection, careful control of the collection procedure ensured that there was no increased aggregation during preparation of radiolabelled platelets, which nevertheless retained function

throughout the measurement period (as determined by aggregation in response to ADP (43 : data not shown). This obviated the need to add inhibitors of aggregation such as prostacyclin during processing. The geometry of the system was designed to maximise the contact of blood with the test surface: the surface area: volume ratio was 23.2 for the test surface and 1.4 for the syringe components, thereby minimising interference from the latter surfaces. Using this system, treatment of the surface with anti-HSA antibody clearly reduced the level of platelet attachment as shown in Figure 8. It is interesting to note that antibody treatment reduced platelet attachment to the same level as observed for the control tubing (Pellethane 55D). As such, the unmodified Carbothane C95A appears to bind more platelets than the control 55D in this system. The lower figure could either represent a basal attachment level or the limit of the assay's sensitivity but cannot be determined here: ideally this would be done by simultaneous microscopic examination. Nonetheless, it is clear that treatment with anti- HSA Mab resulted in reduced platelet attachment.

Cell Attachment to Modified Surfaces The studies presented to date have concentrated on the attachment of haemostatically-relevant platelets. However, surface passivation by HSA implies that the effect would not be limited to platelets. To examine this further, the attachment of a mammalian cell line was investigated. Figure 9 shows that patterning tissue culture wells with specific Mabs enabled control of the initial attachment of BHK cells in the presence of serum. Anti-Fn Mab23 promoted cell attachment by adsorption of Fn from the serum used in the culture medium, to which cells bind via their a5b1 integrin receptors.

Conversely, anti-HSA Mab inhibited cell attachment. A similar effect was observed when glass coverslips were patterned with antibodies. Under the microscope the boundary between attached and non-adherent cells was clearly delineated (Figure 10). The few cells which were attached within the anti-HSA region appeared rounded and poorly spread, implying that they were not tightly bound to the surface. Serum contains several components capable of promoting cell attachment 45, including vitronectin which is distinguished by its ability to adsorb to substrates such as TCPS and glass in the face of competition from other serum components 46, 47. In this system, then, the activity of the surface-bound albumin must be sufficient to overcome the effects of surface-active adhesive components such as Vn.

Fibrinogen Adsorption from plasma onto Modified Surfaces One possible mechanism for the prevention of cell adhesion by the presence of albumin on the surface is the reduction of fibrinogen adsorption.

Cells and platelets can bind fibrinogen via the gpIIb/IIIa (asb1) integrin receptor44, and it is cleaved by thrombin to form the insoluble fibrin scaffold for thrombus formation, giving this protein a central role in platelet adhesion and haemostasis 48. It is thus apparent that preventing fibrinogen adsorption may not only reduce fibrinogen-mediated platelet attachment but also decrease the likelihood of fibrin formation and thrombosis at the surface 48 49. There is a considerable body of literature which shows that adsorbed fibrinogen may be displaced from a surface by competing plasma components, as demonstrated by a characteristic peak in fibrinogen adsorption at intermediate plasma dilutions. This phenomenon is known as the'Vroman effect'and corresponds to that ratio of fibrinogen to competing species which optimises fibrinogen asdorption 35.

In the current study the Vroman effect was evident on the TCPS surface and corresponded to a plasma concentration of approximately 0.3% (Figure 11). Glass, which like TCPS is a hydrophilic surface, has been shown to have a Vroman peak around 0.75% plasma 50 though it is accepted that this value can vary widely depending on the experimental conditions employed 35. Coating the TCPS with the control antibody (anti-EGF; C11) showed a small reduction in the amount of fibrinogen adsorbed. This was probably due to the presence of immunoglobulin G (IgG) molecules on the surface blocking the subsequent adsorption of plasma proteins in a non- specific manner 51. The presence of the anti-HSA Mab on the surface resulted in a profound decrease in adsorption of fibrinogen from plasma. The Vroman peak was still evident on TCPS but was greatly reduced. It is clear from the above that the presence of an albumin-binding antibody on a surface reduced fibrinogen adsorption to levels much lower than would be predicted from the simple non-specific'blocking'effect of IgG molecules on the surface. This is clearly of great benefit in blood-contacting applications.

The mechanism of this albumin-enhanced decrease in fibrinogen adsorption cannot be deduced from this data. In the current application mass action effects will be significant since albumin is the most abundant plasma protein 52. On an albumin-binding surface this will be a major feature, since the probability of albumin being the first molecule to impinge on an unoccupied

binding site is at least seven times higher than that of any other protein Given the relative molecular sizes of albumin (prolate elipsoid, 14nm x 4nm and thickness 4nm 53) and IgG (elipsoid, radius 7.5nm and thickness 3. Onm 54), it is just conceivable that albumin molecules could form an even monolayer over the surface, thereby blocking fibrinogen adsorption. This assumes, however, that the Mab molecules will be homogenously distributed across the surface, and that all of them will be fully competent to bind antigens. Previous evidence indicates that both these assumptions are unlikely to be met, implying that simple direct steric effects are unlikely to account for the dramatic reduction in fibrinogen adsorption. A more rigorous examination might involve studies using antibodies to another'bland'plasma protein, and studies on albumin immobilised via Mabs, including measurement of surface density and distribution, physico-chemical investigations and analysis of the interaction with fibrinogen in solution.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

References 1. J. Brash, T. Horbett,"Proteins at Interfaces: An Overview,"in Proteins at Interfaces II. (1995). T. Horbett and J. Brash (Eds). American Chemical Society, Washington, 1-23.

2. R. Williams, J. Hunt, P. Tengvall,"Fibroblast adhesion onto methyl- silica gradients with and without preadsorbed protein.,"J Biomed Mater Res, 29 (12), 1545-55 (1995).

3. M. Taborelli, L. Eng, P. Descouts, J. Ranieri, R. Bellamkonda, P.

Aebischer,"Bovine serum albumin conformation on methyl and amine functionalized surfaces compared by scanning force microscopy.,"J Biomed Mater Res, 29 (6), 707-14 (1995).

4. J. Lawler, H. Slayter, J. Coligan,"Isolation and characterization of a high molecular weight glycoprotein from human blood platelets.,"J Biol Chem, 253 (23), 8609-16 (1978).

5. P. Underwood,"Mapping of antigenic changes in the haemagglutinin of Hong Kong influenza (H3N2) strains using a large panel of monoclonal antibodies.,"J Gen Virol, 62 (Pt 1), 153-69 (1982).

6. P. Underwood,"Practical considerations of the ability of monoclonal antibodies to detect antigenic differences between closely related variants.," J Immunol Methods, 85 (2), 309-23 (1985).

7. P. Underwood, P. Bean,"Hazards of the limiting-dilution method of cloning hybridomas.,"J Immunol Methods, 107 (1), 119-28 (1988).

8. P. Underwood, J. Kelly, D. Harman, H. MacMillan,"Use of protein A to remove immunoglobulins from serum in hybridoma culture media.,"J Immunol Methods, 60 (1-2), 33-45 (1983).

9. R. Chatelier, T. Gengenbach, Z. Vassic, H. Griesser,"Covalent attachment and non-specific binding of reactive probe molecules onto surfaces,"J Biomater Sci Polym Edn, 7,601 (1995).

10. C. Stahli, V. Miggiano, J. Stocker, T. Staehelin, P. Haring, B. Takacs, "Distinction of epitopes by monoclonal antibodies.,"Methods Enzymol, 92, 242-53 (1983).

11. J. Tresca, R. Ricoux, M. Pontet, R. Engler,"Comparative activity of peroxidase-antibody conjugates with periodate and glutaraldehyde coupling according to an enzyme immunoassay.,"Ann Biol Clin (Paris), 53 (4), 227-31 (1995).

12. E. Fischer, A. Stingl, C. Kirkpatrick,"Migration assay for endothelial cells in multiwells. Application to studies on the effect of opioids.,"J Immunol Methods, 128 (2), 235-9 (1990).

13. N. James, K. Schindhelm, P. Slowiaczek, B. Milthorpe, N. Dudman, G. Johnson, J. Steele,"Endothelial cell seeding of small diameter vascular grafts.,"Artif Organs, 14 (5), 355-60 (1990).

14. P. Underwood,"Problems and pitfalls with measurement of antibody affinity using solid phase binding in the ELISA.,"J Immunol Methods, 164 (1), 119-30 (1993).

15. J. Ara, R. Ali,"Antigenic specificity of anti-ROS DNA antibodies: involvement of lysyl residues in antigen binding.,"Biochem Mol Biol Int, 35 (1), 213-22 (1995).

16. T. Nikula, M. Bocchia, M. Curcio, G. Sgouros, Y. Ma, R. Finn, D.

Scheinberg,"Impact of the high tyrosine fraction in complementarity determining regions: measured and predicted effects of radioiodination on IgG immunoreactivity.,"Mol Immunol, 32 (12), 865-72 (1995).

17. E. Padlan,"Anatomy of the antibody molecule.,"Mol Immunol, 31 (3), 169-217 (1994).

18. D. Gretch, M. Suter, M. Stinski,"The use of biotinylated monoclonal antibodies and streptavidin affinity chromatography to isolate herpesvirus hydrophobic proteins or glycoproteins.,"Anal Biochem, 163 (1), 270-7 (1987).

19. L. Pugliese, A. Coda, M. Malcovati, M. Bolognesi,"Three-dimensional structure of the tetragonal crystal form of egg-white avidin in its functional complex with biotin at 2.7 A resolution.,"J Mol Biol, 231 (3), 698-710 (1993).

20. T. Updyke, G. Nicolson,"Immunoaffinity isolation of membrane antigens with biotinylated monoclonal antibodies and immobilized streptavidin matrices.,"J Immunol Methods, 73 (1), 83-95 (1984).

21. B. C. Cook, G. S. Retzinger,"Elution of Fibrinogen and other plasma proteins from Unmodified and from Letithin-Coated Polystyrene- Divinylbenzene Beads,"J. Colloid. Interface. Sci., 153 (1), 1-12 (1992).

22. W. Norde, C. A. Haynes,"Reversibility and the Mechanism of Protein Adsorption,"in Proteins at Interfaces II. (1995). T. Horbett and J. Brash (Eds). American Chemical Society, Washington, 26-40.

23. R. Honzatko, R. Williams,"Raman spectroscopy of avidin: secondary structure, disulfide conformation, and the environment of tyrosine.," Biochemistry, 21 (24), 6201-5 (1982).

24. G. Dimitriadis,"Effect of detergents on antibody-antigen interaction.," Anal Biochem, 98 (2), 445-51 (1979).

25. J. Potts, I. Campbell,"Fibronectin structure and assembly.,"Curr Opin Cell Biol, 6 (5), 648-55 (1994).

26. M. Suter, J. Butler,"The immunochemistry of sandwich ELISAs. II. A novel system prevents the denaturation of capture antibodies.,"Immunol Lett, 13 (6), 313-6 (1986).

27. K. Winters, J. Walsh, B. Rubin, S. Santoro,"Platelet interactions with fibronectin: divalent cation-independent platelet adhesion to the gelatin- binding domain of fibronectin.,"Blood, 81 (7), 1778-86 (1993).

28. S. Goodman, S. Cooper, R. Albrecht,"Integrin receptors and platelet adhesion to synthetic surfaces.,"J Biomed Mater Res, 27 (5), 683-95 (1993).

29. D. Blockmans, H. Deckmyn, J. Vermylen,"Platelet activation.,"Blood Rev, 9 (3), 143-56 (1995).

30. J. Loftus, J. Choate, R. Albrecht,"Platelet activation and cytoskeletal reorganization: high voltage electron microscopic examination of intact and Triton-extracted whole mounts.,"J Cell Biol, 98 (6), 2019-25 (1984).

31. C. Haycox, B. Ratner,"In vitro platelet interactions in whole human blood exposed to biomaterial surfaces: insights on blood compatibility.,"J Biomed Mater Res, 27 (9), 1181-93 (1993).

32. M. Packham,"The behavior of platelets at foreign surfaces.,"Proc Soc Exp Biol Med, 189 (3), 261-74 (1988).

33. S. Alaeddine, H. Nygren,"Logarithmic growth of Protein Films,"in Proteins at Interfaces II. (1995). T. Horbett and J. Brash (Eds). American Chemical Society, Washington, 41-51.

34. D. Absolom, C. Thomson, L. Hawthorn, W. Zingg, A. Neumann, "Kinetics of cell adhesion to polymer surfaces.,"} Biomed Mater Res, 22 (3), 215-29 (1988).

35. S. Slack, T. Horbett,"The Vroman Effect : A Critical Review,"in Proteins at Interfaces II. (1995). T. Horbett and J. Brash (Eds). American Chemical Society, Washington, 112-128.

36. A. Poggi, M. Stella, M. Donati,"The importance of blood cell-vessel wall interactions in tumour metastasis.,"Baillieres Clin Haematol, 6 (3), 731- 52 (1993).

37. T. Alkhamis, R. Beissinger,"Surface and bulk effects on platelet adhesion and aggregation during simple (laminar) shear flow of whole blood.,"J Biomater Sci Polym Ed, 6 (4), 343-58 (1994).

38. H. Roald, R. Barstad, I. Bakken, B. Roald, T. Lyberg, K. Sakariassen, "Initial interactions of platelets and plasma proteins in flowing non- anticoagulated human blood with the artificial surfaces Dacron and PTFE.," Blood Coagul Fibrinolysis, 5 (3), 355-63 (1994).

39. J. Olson, A. Zaleski, D. Herrmann, P. Flood,"Adhesion of platelets to purified solid-phase von Willebrand factor: effects of wall shear rate, ADP, thrombin, and ristocetin.,"J Lab Clin Med, 114 (1), 6-18 (1989).

40. F. Datz,"Radiolabeled leukocytes and platelets.,"Invest Radiol, 21 (3), 191-200 (1986).

41. J. McCue, D. Stevens, V. Kermon, M. Ashe, L. Heim,"Platelet lesion of collection.,"Scand J Haematol, 25 (4), 301-7 (1980).

42. M. Mant, C. Kappagoda, R. Taylor, J. Quinlan,"Platelet activation caused by cardiac catheter blood collection, and its prevention.,"Thromb Res, 33 (2), 177-87 (1984).

43. G. Rao,"Physiology of blood platelet activation.,"Indian J Physiol Pharmacol, 37 (4), 263-75 (1993).

44. R. Hynes,"Integrins: versatility, modulation, and signaling in cell adhesion.,"Cell, 69 (1), 11-25 (1992).

45. L. Culp, J. Buniel,"Substrate-attached serum and cell proteins in adhesion of mouse fibroblasts.,"J Cell Physiol, 88 (1), 89-106 (1976).

46. M. D. Bale, L. A. Wohlfahrt, D. F. Mosher, B. Tomasini, R. C. Sutton, "Identification of Vitronectin as a Major Plasma Protein Adsorbed on Polymer Surfaces of Different Copolymer Composition.,"Blood, 74 (8), 2698-2706 (1989).

47. J. Steele, G. Johnson, P. Underwood,"Role of serum vitronectin and fibronectin in adhesion of fibroblasts following seeding onto tissue culture polystyrene.,"J Biomed Mater Res, 26 (7), 861-84 (1992).

48. B. Blomback,"Fibrinogen and fibrin--proteins with complex roles in hemostasis and thrombosis.,"Thromb Res, 83 (1), 1-75 (1996).

49. M. Packham,"Role of platelets in thrombosis and hemostasis.,"Can J Physiol Pharmacol, 72 (3), 278-84 (1994).

50. J. Brash, H. P. ten,"Effect of plasma dilution on adsorption of fibrinogen to solid surfaces.,"Thromb Haemost, 51 (3), 326-30 (1984).

51. R. J. Vogt, D. Phillips, L. Henderson, W. Whitfield, F. Spierto, "Quantitative differences among various proteins as blocking agents for ELISA microtiter plates.,"J Immunol Methods, 101 (1), 43-50 (1987).

52. J. Andrade, V. Hlady,"Plasma protein adsorption : the big twelve.," Ann N Y Acad Sci, 516,158-72 (1987).

53. J. Lee, H. Lee,"A wettability gradient as a tool to study protein adsorption and cell adhesion on polymer surfaces.,"J Biomater Sci Polym Ed, 4 (5), 467-81 (1993).

54. P. Esser,"Principles in adsorption to polystyrene.,"Nunc Bulletin., 6, 1-5 (1988).