Journal of Biomaterials Science, Polymer Edition
ISSN: 0920-5063 (Print) 1568-5624 (Online) Journal homepage: http://www.tandfonline.com/loi/tbsp20
Protein adsorption from plasma onto poly(n-alkyl methacrylate) surfaces H.S. Van Damme , T. Beugeling , M.T. Ratering & J. Feijen To cite this article: H.S. Van Damme , T. Beugeling , M.T. Ratering & J. Feijen (1992) Protein adsorption from plasma onto poly(n-alkyl methacrylate) surfaces, Journal of Biomaterials Science, Polymer Edition, 3:1, 69-84, DOI: 10.1163/156856292X00088 To link to this article: http://dx.doi.org/10.1163/156856292X00088
Published online: 02 Apr 2012.
Submit your article to this journal
Article views: 6
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tbsp20 Download by: [University of California Santa Barbara]
Date: 19 June 2016, At: 23:16
Protein
adsorption
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
poly(n-alkyl
from
methacrylate)
onto plasma surfaces
M. T. RATERING and J. FEIJEN H. S. VAN DAMME*, T. BEUGELING, Department of Chemical Technology,Universityof Twente, P.O. Box 217, 7500AE Enschede, The Netherlands Received21 September 1990;accepted 1 February 1991 Abstract-Protein adsorption of human serum albumin (HSA), human fibrinogen (Fg), human immunoglobulinG (IgG), high densitylipoprotein (HDL) and high molecularweightkininogen(HMWK) from plasma onto poly (n-alkyl methacrylate)(PAMA) surfaces was measured using a semi-quantitative enzyme-immunoassay.Adsorption was investigate for PAMA(n = 1) (n is the number of C-atomsin the n-alkyl side chain), PAMA(n = 8) and PAMA(n = 18). PAMA(n = 1) has a relatively hydrophilic surfaceas compared to the more hydrophobicPAMA(n = 8) surface. Both polymershave surface chains which do not reorient after contact with water. The PAMA(n = 18)surface is relativelyhydrophobicbut in this case polymer surface chains and segmentsare able to reorient after contact with water. Protein adsorption was measuredboth as a function of time and as a function of the plasma dilution. If adsorption from plasma was measured as a function of time no exchange of proteins could be observed. The amount of adsorbed protein was always larger in the case of the hydrophobic PAMA(n = 8) as compared to PAMAS(n = 1 and 18), probably due to hydrophobic interactions between the proteins and the PAMA(n = 8) surface. At high plasma concentration relatively large amounts of HDL adsorb onto PAMA(n = 8), indicatingthat this lipoprotein preferentiallyadsorbs onto this surface. Key words: Enzyme-immunoassay; protein adsorptic lipoprotein; hydrophobicity; plasma.
Vroman effect; surface mobility; high density
INTRODUCTION The exchange of proteins adsorbed at solid-liquid interiaces with proteins in solution as a function of time is known as the 'Vroman' effect [1-3]. When a solid surface is exposed to plasma, adsorption of proteins which have a high diffusion coefficient and which are present in high concentrations (e.g., human serum albumin, HSA) will adsorb first. In time HSA will be replaced by larger proteins like G fibrinogen (Fg) or immunoglobulin [4, 5], which may develop a stronger interaction with the surface than albumin. ',, hereafter Fg and IgG may be exchanged by even larger proteins like high molecular weight kininogen (HMWK) [6], or high density lipoprotein (HDL) [7]. The 'Vroman' effect is not only observed as a maximum in the amount of adsorbed protein as a function of time but also as a maximum in the amount of adsorbed protein as a function of the plasma dilution [8]. Brash et al. [9] observed a maximum in the amount of Fg adsorbed onto glass, silanized glass and polyethylene when they measured the adsorption after 5 min as a function of the plasma
70
"
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
Figure 1. Repeating unit of poly(n-alkylmethacrylate).
dilution. At low plasma concentrations the Fg concentration is also low and therefore the adsorbed amount of Fg will be low. At a high plasma concentration, `he concentrations of the proteins are relatively high, leading to a fast exchange of adsorbed Fg molecules with other protein molecules in solution, for example HMWK. At intermediate plasma concentrations a relatively large amount of Fg may adsorb which will be slowly displaced by other proteins or will be irreversibly adsorbed. Because the 'Vroman' effect involves exchanges of various adsorbed proteins with proteins from a protein solution or plasma, it is expected that this effect is highly dependent on the properties of the material surface. We have previously reported on the surface mobility and structural transitions of (PAMAs Fig. 1) measured by dynamic contact angle poly(n-alkyl methacrylates) measurements [ 10] . The receding angle of PAMAs as a function of side chain length (n) measured at room temperature is gradually increasing from n = 1 to 10. At longer chain lengths (n = 12-18) a sharp decrease in the receding angle is observed indicating that reorientation of side chains takes place which results in relatively hydrophilic surfaces in an aqueous environment. For the adsorption studies of proteins adsorbing from plasma we have selected PAMAs(n = 1, 8 and 18). PAMA(n = 1) is the most hydrophilic polymer of the three PAMAs because it has a small hydrophobic n-alkyl group compared to = = a higher advancing and 8 and exhibited both 8) PAMAs(n 18). PAMA(n = to the relatively long contact to due PAMA(n 1) receding angle compared = For tail of the ester hydrophobic PAMA(n 18) the highest advancing group. contact angle of the three PAMAs was observed due to the long hydrophobic octadecyl group. However, the receding contact angle observed for PAMA(n = 18) was lower than the receding angles observed for the two other PAMAs. The low receding contact angle measured for PAMA(n = 18) could be explained by a reorientation of surface groups and polymer segments after contact of the polymer surface with water. A similar reorientation is unlikely in the case of PAMAs(n = 1 and 8) because the surface groups and segments of these polymers are in a rigid state while those of PAMA(n = 18) are in a mobile state. The aim of this work was to investigate the adsorption and exchange behaviour of proteins adsorbing from plasma onto PAMAs(n = 1, 8 and 18) as a function of time and plasma concentration (Vroman effect) and to relate the results to the surface characteristics of the different PAMAs used. The proteins studied were HSA, Fg, IgG, HMWK and HDL. A two step enzyme-immunoassay (EIA) was used to of the proteins adsorbed to the different measure relative surface concentrations surfaces.
71 MATERIALSAND METHODS
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
Polymer surfaces PAMAs(n = 1, 8 and 18), were synthesized by radical Poly(n-alkyl methacrylates), polymerization of the corresponding n-alkyl methacrylates as described before (10). Polymer surfaces were obtained by dip-coating clean glass coverslips (1.8 x 1.8 cm, Chance Propper LTD, Swethwick, Warley, UK) in 3 070(w/v) solutions of PAMA in toluene P.A. (Merck, Darmstadt, FRG). The coating was dried overnight in vacuo at 50-60°C. To prevent detachment of the very hydrophobic PAMA(n = 8) from the glass slides during immersion in water, glass slides were pretreated with a solution of trimethylchlorosilane (7 v/vO7o, Merck) and pyridine P.A.(7 v/vO7o, Merck) in toluene at room temperature for 1 h. The silanized glass slides were first rinsed with ethanol P.A. (Merck) and ultrafiltrated water and then dried at 110°C in vacuo for 1 h. The slides were then dip-coated with PAMA(n = 8) as described before. Characterization
of surfaces
Polymer surfaces were characterized by dynamic contact angle measurements using a Wilhelmy plate technique as described in ref. [10]. Water contact angles of the polymer surfaces were measured at 20°C. Protein
adsorption
from plasma
In order to measure the adsorption of proteins from diluted human plasma to polymer coated glass slides, a two step enzyme-immunoassay (EIA) was used of al. Breemhaar et according to the method [7] and modified by Poot [11]. The principle of the method is illustrated in Fig. 2. The coated glass slides were mounted in a special test device. The test device enables 24 adsorption experiments with two polymer coated glass slides. The surface area of the polymer coating in each well of the test device is 0.9 cm2. The total volume of each well is approximately 800,ul. In order to prevent the formation of an air-liquid-solid interface upon contact of with 24 the coated the wells were filled with 200,u1 of plasma (solution) glass, buffered saline Darmstadt, (PBS, NaCl(Merck, FRG) 0.140 phosphate ImmediNa2HP04(Merck) NaH2po4 (Merck) 1.3.10-3 mol l-1). thereafter the was started of adsorption experiment ately by adding 200 ,ul (diluted)
Figure 2. Principle of the enzyme-immunoassay(EIA) for detection of adsorbed proteins.
72
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
human
plasma (single donor, Bloedbank, Twente-Achterhoek, Enschede, The Netherlands). During pipetting of the plasma (solution), the end of the pipettor tip did not touch the test surface and was kept under the liquid surface. After gently mixing using the pipettor, the wells were covered with tape to avoid evaporation of water. After the desired adsorption time the wells were rinsed several times with PBS containing 0.005OJo(w/v) Tween-20 (Sigma, St. Louis, USA). The first step of the EIA was started by adding 200,ul of the first antibody solution (1%(v/v)). In the case of HSA, Fg and IgG the first antibodies were produced by rabbits and obtained from the Central Laboratory of the Netherlands Red Cross Blood Transfusion Service (CLB, Amsterdam, The Netherlands). In the case of HDL the first antibody was also a rabbit antibody (Behringwerke AG, Marburg, FRG). The first antibody directed against HMWK was a goat antibody (Dept. of Haematology, University After an exposure time of 1 h the first antiHospital, Utrecht, The Netherlands). body solution was removed and the wells were rinsed with PBS containing Tween-20. Then 200//1 of a solution of the enzyme-labeled second 0.005%(w/v) antibody (conjugate) (United States Biochemical Co, Cleveland, USA) was added. This second antibody was either sheep anti-rabbit IgG (200 times diluted) or rabbit anti-goat IgG (6000 times diluted) depending on the type of first antibody. All first and second antibodies were polyclonal. The bound enzyme was horse-radish peroxidase. After 1 h the conjugate solution was removed. After rinsing with PBS-Tween20 solution, 200,u1 of a solution of the leuko dye (0.1 g/1, 3,3,5,5-tetramethylbenzidine) ((Fluka AG, Buchs, Switzerland) and the enzyme substrate (3olo(v/v) hydrogen peroxide (Merck)) was pipetted into the wells. The reaction was terminated after 30 min by addition of 100 ,ul 2 M sulfuric acid (Merck) and the absorbance of the dye solution was measured at 450 nm (Reader Micro Elisa System, Organon Teknika, Turnhout, Belgium). In order to minimize possible errors in the results of experiments in which protein adsorption was measured as a function of time, these experiments were started at different moments in order to add the first antibody solution into all wells at the same moment. RESULTS Polymer
surfaces
The results of the characterization of the polymer surfaces by dynamic contact angle measurements are summarized in Table 1. The advancing contact angles are higher for PAMAs with longer n-alkyl side chains. After contact with water, surface groups and segments of PAMA(n = 18) are able to reorient. The reorientation results in a lower receding contact angle in the case of PAMA(n = 18) (48°) as compared to the receding contact angle observed for PAMAs(n = 1 (65°) and 8 (73°)). Protein
adsorption
from plasma
Because the results of the EIA experiments showed a day to day fluctuation, in each experiment two materials were compared using PAMA(n = 1) as a reference material. All absorbance values given in the figures are the averages of data obtained
73
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
Table 1. Water contact angles of PAMA surfaces at T = 20°C.
s.d.: standard deviation from four independent experiments, which were carried out simultaneously. Absorbance values presented in the figures were considered to be significantly different when these values differed more than the sum of their standard deviations. Adsorption as a function of time. In experiments in which protein adsorption was measured as a function of time (15 s-20 h), two plasma dilutions (1 :1and 1 : 1000) were used. The adsorption of the following proteins was studied: albumin (HSA), G (IgG), high density lipoprotein (HDL) and high fibrinogen (Fg), immunoglobulin molecular weight kininogen (HMWK).
Figure 3. Adsorption kinetics of proteins adsorbed from 1:1diluted plasma onto PAMAs(n = 1 and 8). Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) IgG adsorption; (d) HDL adsorption.
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
74
Figure4. Adsorption kineticsof proteins adsorbed from 1:1diluted plasma onto PAMAS(n = 1 and 18). Each absorbance value (A450)is the average value obtained from four experiments.Bars indicate standard deviations. (a) HSA adsorption; (b) Fb adsorption; (c) IgG adsorption; (d) HDL adsorption. In Fig. 3 the amounts of protein adsorbed from 1:11 diluted plasma onto PAMAs(n = 1 and 8) surfaces are compared. For HSA and HDL a significantly higher adsorption onto PAMA(n = 8) as compared to PAMA(n = 1) was found for all exposure times. At long exposure times ( - 12 h) the IgG surface concentration on PAMA(n = 1) equals the one on PAMA(n = 8), while the surface concentration of = is 17 min. on after t = Fg only higher PAMA(n 8) A significant increase of the amount of HDL as a function of time is found for on PAMA(n = 1) an increase of the amount of IgG PAMA(n = 8). Furthermore, at long exposure times is observed. In Fig. 4 the results of experiments with regard to the adsorption of proteins from 1:1diluted plasma onto PAMAs(n = 1 and 18) are shown. Differences between the amounts of adsorbed protein to PAMA(n = 1) and onto PAMA(n = 18) were only significant in the case of IgG. The amounts of adsorbed IgG were always larger in the case of PAMA(n = 18) as compared to PAMA(n = 1 ) except for the longest adsorption time where no significant difference was observed. After longer adsorption times the amounts of HSA and HDL adsorbed to both polymers increased 1 and 18) as a slightly. No change in the amounts of Fg adsorbed to PAMAs(n = function of time could be observed. In Fig. 5 the adsorption of HSA, Fg, IgG and HDL from 1:1000 diluted plasma onto PAMAs(n = 1 and 8) is shown. The amounts of HSA and Fg adsorbed onto PAMA(n = 1 ) increased as a function of time, while no significant change in the amounts of adsorbed HSA and Fg to PAMA(n = 8) could be observed. No change in the amount of IgG adsorbed to PAMA(n = 8) was found, while the amount
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
75
diluted plasma onto PAMAs(n = 1 and Figure 5. Adsorption kineticsof proteins adsorbed from 1 : 1000 8). Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) IgG adsorption; (d) HDL adsorption.
Figure 6. Adsorption kineticsof proteins adsorbed from 1:1000diluted plasma onto PAMAs(n = 1 and 18). Each absorbance value (A450)is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) IgG adsorption; (d) HDL adsorption.
76
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
adsorbed to PAMA(n = 1) varied as a function of time. The amount of adsorbed HDL to PAMA(n = 8) increased significantly as a function of time and reached a high value compared to the values observed in the case of PAMA(n = 1). In Fig. 6 the amounts of adsorbed HSA, Fg, IgG and HDL from 1 : 1000 diluted plasma to PAMAs(n = 1 and 18) are compared. Only the differences between the amounts of adsorbed Fg to these surfaces were significant. The amount of Fg adsorbed to PAMA(n = 1) was lower at short exposure times and increased somewhat faster than the amount of Fg adsorbed to PAMA(n = 18). For both PAMAs a minimum in the adsorbed amount of IgG as a function of the adsorption time was found. Adsorption as a function of plasma dilution. The adsorption time was 1 h for experiments in which protein adsorption was measured as a function of the plasma dilution (1:1 to 1:100000). In a series of experiments carried out in 1 day the amounts of protein adsorbed from plasma solutions to two different PAMA surfaces were measured. In Fig. 7 the adsorption of HSA, Fg, HMWK and HDL from plasma solutions is given for PAMAs(n = 1 and 8). Significant differences between the amounts of HSA and HDL adsorbed to PAMAs(n = 1 and 8) were observed. In both cases the amounts of adsorbed protein to PAMA(n = 8) were higher than those adsorbed to At intermediate plasma dilutions the PAMA(n = 1) at high plasma concentrations. amount of Fg adsorbed to PAMA(n = 1) was slightly higher than the amount
Figure 7. Protein adsorption from plasma to PAMAs(n = 1 and 8) as a function of the plasma dilution. The adsorption time was 1 h. Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) HMWK adsorption; (d) HDL adsorption.
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
77
Figure8. Protein adsorption from plasma to PAMAs(n = 1 and 18)as a function of the plasma dilution. The adsorption time was 1 h. Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) HMWK adsorption; (d) HDL adsorption. adsorbed to PAMA(n = 8). For both Fg and HMWK a maximum in the amount of adsorbed protein was found at intermediate plasma dilution. In Fig. 8 the results of similar adsorption are shown for experiments = 1 and Almost no were differences observed between PAMAs(n 18). significant the amounts of protein adsorbed to PAMAs(n = 1 and 18). For HSA, Fg and HMWK a maximum in the adsorbed amount of protein was found at intermediate plasma dilution (1:10,000 to 1:1000). In the case of HDL an adsorption plateau was present at high plasma concentrations. In Fig. 9 protein adsorption of PAMA(n = 8) is compared with the adsorption to PAMA(n = 18). Adsorption of HSA and HDL to PAMA(n = 8) was significantly to higher compared with PAMA(n = 18). On the other hand Fg adsorption = was in Also these maxima in the amounts of PAMA(n 18) higher. experiments adsorbed Fg and HMWK were found at intermediate plasma dilution. DISCUSSION In spite of numerous studies, the exact composition of the protein layer adsorbed to a polymer surface from a multicomponent mixture such as plasma is not yet known. The reason for this is that it is extremely difficult to quantitatively detect one specific protein, often present in low surface concentration, among many others which are adsorbed. Promising techniques to study protein adsorption from complex mixtures are In this investigation a two step enzyme-immunoassay enzyme-immunoassays.
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
78
Figure 9. Protein adsorption from plasma to PAMAs(n = 8 and 18)as a function of the plasma dilution. The adsorption time was 1 h. Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) HMWK adsorption; (d) HDL adsorption.
[7, 11, 24] has been used to detect the adsorption of HSA, IgG, Fg, HDL and HMWK from human plasma to PAMA surfaces. The advantage of a two step is that usually only one enzyme-labeled enzyme-immunoassay antibody is used, whereas in the case of a one step assay many enzyme-labeled antibodies are necessary. The relative adsorption data are expressed as absorbances of the generated dye in the enzyme-immunoassays. These data are not directly related to amounts of adsorbed Protein molecules conformational protein/cm2. generally undergo and after As a several changes during adsorption. consequence antigenic determinants of a protein molecule may lose their specific structure and may not be able to react with binding sites of the applied antibody. This phenomenon not only material on the and the surface but also on the depends protein possibility of the the is time unfold and which to surface, protein dependent and spread upon influenced by the surface concentration of proteins already adsorbed. Moreover, in the case of multiple layer adsorption, sublayers cannot be detected by the enzymeimmunoassays. Therefore a comparison of the data obtained by this technique with those of a quantitative method like radiolabeling is very difficult. When polymer surfaces are exposed to (diluted) plasma, the composition of the diffusion rates of initially adsorbed protein layer depends on protein concentrations, the proteins and free energies of adsorption. It is expected that plasma proteins which are present in a high concentration and which are small adsorb first. In course of time an exchange of the adsorbed proteins with more strongly adsorbing proteins
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
79 may take place (Vroman effect). Vroman and Adams [12] found that proteins which adsorb from plasma onto glass, displace each other in the following sequence: albumin, IgG, fibrinogen, fibronectin, HMWK and factor XII; i.e. from the most concentrated to the least concentrated protein. In the experiments in which protein adsorption from 1:1 diluted plasma to the PAMAs was studied as a function of time, no significant maxima in the amounts of adsorbed HSA, Fg, IgG and HDL were detected (Figs 3 and 4). The shortest time of exposure of the PAMAs to plasma that could be realized with the EIA technique was 15 s. Probably, an exchange of the adsorbed proteins with proteins in plasma had already taken place within 15 s and had resulted in an adsorbed protein layer in which a displacement of proteins was (practically) impossible. Only the amount of HDL adsorbed to PAMA(n = 8) increased significantly after 20 min of exposure. Probably HDL adsorbs irreversibly to PAMA(n = 8), and not to PAMAs(n = 1 and 18) because PAMA(n = 8) is more hydrophobic than the other two (Table 1). For both series it was found that after long exposure times the adsorption of IgG onto PAMA(n = 1) increased considerably. This phenomenon cannot be explained at the moment. An important result from these experiments is the high adsorption of all proteins 1 and 18). A onto PAMA(n = 8) compared with the adsorption onto PAMAs(n = between and has interaction surfaces been observed by proteins strong hydrophobic several investigators [13-18]. In the case of a hydrophobic surface water molecules are ordered in an ice-like structure at the surface and have a much lower entropy than water molecules in the bulk. The interaction between a hydrophobic surface and a protein originates mainly from an entropy gain due to water desorption from the solid surface and from the protein molecule [ 19-22] . The low HSA adsorption from 1:1 diluted plasma to PAMA(n = 1 ) and PAMA(n = 18) as compared with the adsorption to PAMA(n = 8) may be due to the fact that HSA is more reversibly adsorbed to PAMAs(n = 1 and 18) compared with PAMA(n = 8) which has the most hydrophobic surface of the three polymers. The results of protein adsorption from 1:1000 diluted plasma (Figs 5 and 6) to the PAMAs as a function of time are roughly comparable with those obtained with 1 :1 diluted plasma. The most striking difference between the adsorption values obtained with the two plasma dilutions is the high adsorption of HSA onto PAMA(n = 8) with regard to PAMA(n = 1) at a plasma dilution of 1 : (Fig. 3) compared to the HSA adsorption from 1:1000 diluted plasma (Fig. 5). A high Fg adsorption from 1:1000 diluted plasma was observed for all three PAMAs. The high adsorption of Fg from 1:1000 diluted plasma to the PAMAs may be caused by a lower surface concentration of plasma proteins in the beginning of the adsorption process as compared to PAMAs which are exposed to 1:1diluted Hence plasma. Fg molecules have the opportunity to spread out or unfold, resulting in irreversible protein adsorption [23]. As a consequence, adsorbed Fg molecules will hardly be exchanged with proteins present in the plasma solution. In the case of 1 : diluted plasma, the surface will be totally occupied by proteins, and molecules of Fg which are adsorbed, have a minor opportunity to unfold or spread out upon the surface. Such molecules may be exchanged with protein molecules from the plasma solution. It has been shown by Breemhaar et al. [7]. Poot [11]] and Beugeling [23] that HDL is most probably involved in the displacement of proteins adsorbed from plasma to
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
80 hydrophobic polymers like PVC, polystyrene and polyethylene. It is likely that HDL is also involved in the displacement of proteins, adsorbed from (diluted) plasma to the PAMA surfaces. The relatively large amounts of HDL which adsorb from 1 :1 and 1:1000 diluted plasma to PAMA(n = 8) (Figs 3-6) may support this assumption. However, Figs 4 and 6 show rather low adsorption values for HDL on PAMA(n = 1) and PAMA(n = 18). These low values may be due to the fact that only parts of the HDL molecule consist of apoprotein A-1 to which the first antiof antibody body is directed. This results in a relatively low surface concentration which in turn leads to a low absorbance of the molecules, generated dye (A450 value). The minima in the amount of IgG adsorbed from 1:1000 diluted plasma to 1 and 18) as a function of time could not be explained. PAMAs(n = The 'Vroman effect' is not only observed as a maximum in the adsorption of proteins from plasma as a function of time but also as a maximum in the adsorption of time [9]. proteins as a function of the plasma dilution at a fixed adsorption In general, we also found maxima in the amounts of adsorbed protein as a function of plasma dilution. This was observed for HSA adsorbed to PAMAs(n = 1 and 1, 8 and 18) (Figs 7b, 18) (Figs 7a, 8a and 9a), and for Fg adsorbed to PAMAs(n = 8b and 9b) as well as for HMWK adsorbed to PAMAs(n = 1, 8 and 18) (Figs 7c, 8c and 9c). At high plasma dilution the amounts of proteins adsorbed onto a material surface are low. At high plasma concentrations the amount of a particular protein adsorbed to the surface may also be low because the adsorbed protein may have been displaced by other proteins. In several studies in which glass or glass-like surfaces and polymers were exposed to plasma it was demonstrated that the amount of adsorbed Fg was larger if HMWK-deficient was used [2, 6, 11, or blood plasma 24-26]. This indicates that HMWK may play an important role in the exchange of Fg adsorbed to material surfaces. It is expected that surface hydrophilicity/ role in the exchange of adsorbed protein hydrophobicity plays an important molecules with proteins in solution. When proteins are adsorbed more strongly the exchange is expected to be slower. Vroman et al. [2] contacted surfaces with different hydrophilicity/hydrophobicity with human plasma. They observed platelet adhesion onto the hydrophobic surfaces and no platelet adhesion to glass-like surfaces after adsorption of proteins from plasma. Because platelets preferentially adhere to fibrinogen-coated surfaces they concluded that on the glass-like materials, Fg was exchanged with other proteins while on the hydrophobic surface Fg was not exchanged. Wojciechowski et al. [27] studied the transient adsorption of fibrinogen (Vroman effect) using glass, silylated glass and polymer surfaces. None of the measured adsorption could be correlated with the water contact characteristics angles of the different surfaces. Elwing et al. [28] studied the competition between Fg and HMWK during adsorption on different surfaces after incubation with human plasma. These authors used silanized silicon wafers with a wettability gradient over the surface. It was observed that for all locations at the surface a decrease in anti-fibrinogen binding as a function of time was not associated with an increased anti-HMWK binding. This may indicate that other proteins than HMWK are also involved in the Fg displacement. Mulzer and Brash [29] found that protein layers formed on However, cellulose hemodialysis membranes of poly (methyl methacrylate), Cuprophane,
81
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
acetate and saponified cellulose ester during clinical use contained significant amounts of HMWK, while small amounts of adsorbed Fg seemed to be present. It was suggested that Fg adsorbed from the blood to the membranes had been displaced by HMWK. In this respect it must be mentioned that the authors did not investigate the adsorption of HDL to the dialysis membranes. The adsorption studies with respect to the three PAMA surfaces show that the This amounts of adsorbed Fg and HMWK are low at high plasma concentrations. indicates that both Fg and HMWK are exchanged with (a) more strongly adsorbing protein(s). It has been mentioned above that HDL is most probably involved in this exchange process. The course of protein adsorption to the PAMA surfaces as a function of the plasma dilution (Figs 7-9) will now be explained in more detail. At a high plasma dilution the adsorbed amount of protein is either low due to depletion of the solution or due to diffusion limitation. In order to get a rough estimate of the effects of diffusion and depletion on protein adsorption in our experiments, a simplified model has been used. In this model the effects of convection are neglected, although during the initial few seconds plasma (solution) had to be mixed with buffer in the wells. Furthermore it was intended to calculate maximal amounts of adsorbed protein as a function of time, neglecting the fact that all collissions of protein molecules with the surface may not lead to adsorption. The flux of a protein towards a surface can be calculated using Eqn (1).
in which JX - o is the flux of a protein towards a water/polymer interface assuming diffusion of the protein from a semi-infinite medium towards a flat plate. When every protein molecule which collides with the surface adsorbs, Jx == 0 represents the = the concentration maximal adsorption rate at static conditions at c(t = 0, x > 0), c(t 2:: 0, x = 0) = 0 and D is the diffusion coefficient of the protein. The maximal amount of adsorbed protein (rmaX) after a certain time can be calculated from Eqn (2), obtained by integrating Eqn (1).
A reason for the low adsorption values at low plasma concentration may be which is necessary to diffusion limitation. The minimal protein concentration form a monolayer of adsorbed protein in 1 h will be calculated using Eqn (2). Albumin is present in the highest concentration (c = 40 g/1) and has a relatively high diffusion coefficient (D = 6.1 10-' cm2 s-'). The minimal concentration (c) to form a monolayer of side-on HSA molecules (HSA side-on: 0.2 f.1g/cm2) can now be calculated using Eqn (2). It is again assumed that every molecule which collides with the surface adsorbs. This results in c = 3.8 10-3 g/1 or a plasma dilution of 1:10,000. Another possible reason for the low adsorption values could be depletion of the solution which can be expressed in terms of depletion depth. The depletion depth
82
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
(Xp) is defined as the depth where the concentration of the protein is just below 2007o of its original value and can be calculated with Eqn (3) [30].
In the experiments in which protein adsorption was measured as a function of the plasma dilution the adsorption time was always 1 h. Then the depletion depth for a protein with a diffusion coefficient of 4.10-' cm2/s is 0.7 mm independent of the The depletion depth is larger for proteins with higher diffuprotein concentration. sion coefficients. Because the fluid layer in the EIA experiments had a thickness of about 4.5 mm, depletion cannot be the cause of the small amounts of adsorbed proteins at high plasma dilution. From these results it may be concluded that the low amounts of adsorbed protein observed when the plasma is diluted more than 10000 times, are due to diffusion limitation. At such high dilutions the surface is not totally occupied by adsorbed proteins within the first hour of contact but the solution is not depleted. When plasma is diluted less than 10000 times the composition of the adsorbed protein layer after one hour will not only depend on the diffusion rates of the various proteins but also on the competition between these proteins. The maximum in the amount of adsorbed protein observed in the experiments mentioned above is often present at a plasma dilution of about 1:1000. Assuming again that all protein molecules which collide with the surface adsorb, the maximal amount of adsorbed protein on the surface after a particular time can be estimated for each plasma concentration. The results are given in Table 2. The actually adsorbed amounts of protein are expected to be lower than the calculated ones because a collision of a protein molecule with the surface only results in adsorption if the activation energy for adsorption is low and if the interaction forces are sufficiently high [31]. Although maxima in the amounts of adsorbed protein were often observed for adsorption from plasma which was diluted 1000-10000 times, it may be concluded from Table 2 that the adsorbed amounts may still be very low because of the slow diffusion of proteins to the polymer surface. For example in Fig. 7b the amount of adsorbed Fg was maximal for a plasma dilution of 1 : 10000. Based on the model the maximal amount of Fg transported as a result of diffusion towards the surface is only 9.1 ng/cm2. For the other plasma dilutions lower amounts of adsorbed Fg were measured. This suggests that the amounts of Fg adsorbed from diluted or undiluted plasma do not exceed 9.1 ng/cm2 for an adsorption time of one hour. For HMWK this value is even lower: 0.15 ng/cm2. Table 2. Maximal amount of protein adsorbed onto a surface, after 1 h at various plasma dilutions
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
83 Because HSA is a small and relatively hydrophilic protein it was expected that this protein should be easily displaced from the PAMA surfaces by other proteins. This seems to be the case for PAMAs(n = 1 and 18) where low adsorption values for HSA are found at high plasma concentrations (Figs 7a, 8a and 9a). However, plasma solutions to relatively large amounts of HSA adsorb from concentrated PAMA(n = 8) (Figs 7a and 9a). This is also observed in Fig. 4a, which shows the adsorption of HSA from 1 :1diluted plasma to PAMA(n = 8) as a function of time. was earlier explained by assuming that HSA is irreversibly This phenomenon adsorbed to PAMA(n = 8), which has the most hydrophobic surface of the three polymers after contact with an aqueous solution. PAMA(n = 18) with long hydrophobic n-alkyl chains could give rise to strong protein adsorption, but in contrast to PAMAs(n = 1 and 8) surface groups and segments of PAMA(n = 18) can after contact with reorganize in an aqueous solution. Due to this reorganization water a low receding contact angle was observed for PAMA(n = 18) as shown in of the surface groups and Table 1. It may be concluded that the reorganization segments of PAMA(n = 18) after contact with (diluted) plasma results in surface with respect to protein which are comparable to adsorption properties PAMA(n
= 1).
CONCLUSIONS The results of the adsorption experiments in which protein adsorption from plasma to the PAMA surfaces was studied as a function of the plasma dilution suggest diffusion limitation at high plasma dilution. At high plasma concentrations the amounts of HSA, Fg and HMWK adsorbed to the PAMA's are low compared to the amounts adsorbed at intermediate plasma dilution, except for HSA adsorbed to PAMA(n = 8), indicating that these proteins are displaced from the surface by other plasma proteins. In the case of PAMA(n = 8) one of these proteins may be HDL, because at high plasma concentrations higher surface concentrations of HDL adsorbed on PAMA(n = 8) as compared to PAMAs(n = 1 and 18) were observed. Also the experiments, in which protein adsorption from plasma was studied as a function of the adsorption time, show that a strong interaction of HDL with the PAMA(n = 8) surface occurs. Moreover larger amounts of the other proteins adsorbed from plasma solutions to PAMA(n = 8), as compared to PAMAs(n = 1 and 18). There is evidence that the interaction between PAMA(n = 8) and proteins is caused by a strong hydrophobic interaction. Due to reorientation of surface groups and segments of PAMA(n = 18) in an aqueous solution a similar adsorption behaviour of proteins adsorbing from plasma onto PAMAs(n = 18 and 1) is observed. When the adsorption of proteins from plasma was measured as a function of time no decrease in the amounts of adsorbed protein could be observed. This suggests that after the first 15 s of exposure of PAMA surfaces to plasma no displacement of adsorbed proteins takes place. REFERENCES 1. 2. 3. 4.
L. Vroman and A. L. Adams, Surface Sci. 16, 438 (1969). L. Vroman, A. L. Adams, G. C. Fischer, and P. C. Munoz, Blood 55, 156 (1980). L. Vroman, A. L. Adams, J. Colloid Interface Sci. 111, 391 (1986). S. Winters, R. M. Gendreau, R. I. Leininger, and R. J. Jakobsen, Appl. Spectrosc. 36, 404 (1982).
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
84 5. R. M. Gendreau, R. I. Leininger, S. Winters and R. J. Jakobsen, in: Biomaterials: Interfacial Phenomena and Applications, Advances in chemistry series 199, p. 371, S. L. Cooper and N. A. Peppas (Eds). A.C.S., Washington D.C (1982). 6. J. L. Brash and P. ten Hove, J. Biomed. Mater. Res. 23, 157 (1989). 7. W. Breemhaar, E. Brinkman, D. J. Ellens, T. Beugelingand A. Bantjes, Biomaterials5, 269 (1984). 8. T. A. Horbett, Thromb Haemostasis 51, 174 (1984). 9. J. L. Brash and P. ten Hove, Thromb Haemostasis 51, 326 (1984). 10. H. S. van Damme, A. H. Hogt and J. Feijen, J. Colloid Interface Sci. 114, 167 (1986). 11. A. Poot, Thesis, Universityof Twente (1989). 12. I. Vroman and A. L. Adams, in: Proteins at Interfaces. p. 154,J. L. Brash and T. A. Horbett, (Eds). ACS SymposiumSeries 343, A.C.S., Washington D.C. (1987). 13. J. L. Brash and V. J. Davidson, ThrombosisRes. 9, 249 (1976). 14. E. Brynda, N. A. Cepalova and M. Stol, J. Colloid Interface Sci. 18, 685 (1984). 15. H. Y. K. Chuang, W. F. King and R. G. Mason, J. Lab. Clin. Medicine 92, 483 (1978). 16. Y. L. Cheng, S. A. Darst, and C. R. Robertson, J. Colloid Interface Sci. 118, 212 (1987). 17. D. E. Dong, J. D. Andrade and D. L. Coleman, J. Biomed. Mater. Res. 21, 683 (1987). 18. F. Grinnell and M. K. Feld, J. Biomed. Mater. Res. 15, 363 (1981). 19. W. J. Dillman Jr., and I. F. Miller, J. Colloid Interface Sci. 44, 221 (1973). 20. P. G. Koutsoukos, W. Norde and J. Lyklema, J. Colloid Interface Sci, 95, 385 (1983). 21. W. Norde and J. Lyklema, J. Colloid Interface Sci. 66, 295 (1978). 22. W. Norde and J. Lyklema, J. Colloid Interface Sci. 71, 350 (1979). 23. T. Beugeling,in: Modern Aspects of Protein Adsorption. Y. F. Missilis(Ed.). In press. 24. J. L. Brash, C. F. Scott, P. ten Hove, P. Wojciechowskiand R. W. Colman, Blood 71, 932 (1988). 25. A. H. Schmaier, L. Silver, A. L. Adams, G. C. Fischer, P. C. Munoz, L. Vroman and R. W. Colman, ThrombosisRes. 33, 51 (1983). 26. A. Poot, T. Beugeling,W. G. van Aken and A. Bantjes, J. Biomed. Mater. Res. 24, 1021(1990). 27. P. Wojciechowski,P. ten Hove and J. L. Brash, J. Colloid Interface Sci. 111, 455(1986). 28. H. Elwing, A. Askendal and I. Lundstrom, J. Biomed. Mater. Res. 21, 1023(1987). 29. S. R. Mulzer and J. L. Brash, J. Biomed. Mater. Res. 23, 1483(1989). 30. W. J. Beek and K. M. K. Muttzall, Transport Phenomena. John Wiley & Sons Ltd., New York (1977). 31. J. D. Andrade and V. Hlady, Ann. NY Acad. Sci. 516, 158(1987).
ISSN: 0920-5063 (Print) 1568-5624 (Online) Journal homepage: http://www.tandfonline.com/loi/tbsp20
Protein adsorption from plasma onto poly(n-alkyl methacrylate) surfaces H.S. Van Damme , T. Beugeling , M.T. Ratering & J. Feijen To cite this article: H.S. Van Damme , T. Beugeling , M.T. Ratering & J. Feijen (1992) Protein adsorption from plasma onto poly(n-alkyl methacrylate) surfaces, Journal of Biomaterials Science, Polymer Edition, 3:1, 69-84, DOI: 10.1163/156856292X00088 To link to this article: http://dx.doi.org/10.1163/156856292X00088
Published online: 02 Apr 2012.
Submit your article to this journal
Article views: 6
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tbsp20 Download by: [University of California Santa Barbara]
Date: 19 June 2016, At: 23:16
Protein
adsorption
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
poly(n-alkyl
from
methacrylate)
onto plasma surfaces
M. T. RATERING and J. FEIJEN H. S. VAN DAMME*, T. BEUGELING, Department of Chemical Technology,Universityof Twente, P.O. Box 217, 7500AE Enschede, The Netherlands Received21 September 1990;accepted 1 February 1991 Abstract-Protein adsorption of human serum albumin (HSA), human fibrinogen (Fg), human immunoglobulinG (IgG), high densitylipoprotein (HDL) and high molecularweightkininogen(HMWK) from plasma onto poly (n-alkyl methacrylate)(PAMA) surfaces was measured using a semi-quantitative enzyme-immunoassay.Adsorption was investigate for PAMA(n = 1) (n is the number of C-atomsin the n-alkyl side chain), PAMA(n = 8) and PAMA(n = 18). PAMA(n = 1) has a relatively hydrophilic surfaceas compared to the more hydrophobicPAMA(n = 8) surface. Both polymershave surface chains which do not reorient after contact with water. The PAMA(n = 18)surface is relativelyhydrophobicbut in this case polymer surface chains and segmentsare able to reorient after contact with water. Protein adsorption was measuredboth as a function of time and as a function of the plasma dilution. If adsorption from plasma was measured as a function of time no exchange of proteins could be observed. The amount of adsorbed protein was always larger in the case of the hydrophobic PAMA(n = 8) as compared to PAMAS(n = 1 and 18), probably due to hydrophobic interactions between the proteins and the PAMA(n = 8) surface. At high plasma concentration relatively large amounts of HDL adsorb onto PAMA(n = 8), indicatingthat this lipoprotein preferentiallyadsorbs onto this surface. Key words: Enzyme-immunoassay; protein adsorptic lipoprotein; hydrophobicity; plasma.
Vroman effect; surface mobility; high density
INTRODUCTION The exchange of proteins adsorbed at solid-liquid interiaces with proteins in solution as a function of time is known as the 'Vroman' effect [1-3]. When a solid surface is exposed to plasma, adsorption of proteins which have a high diffusion coefficient and which are present in high concentrations (e.g., human serum albumin, HSA) will adsorb first. In time HSA will be replaced by larger proteins like G fibrinogen (Fg) or immunoglobulin [4, 5], which may develop a stronger interaction with the surface than albumin. ',, hereafter Fg and IgG may be exchanged by even larger proteins like high molecular weight kininogen (HMWK) [6], or high density lipoprotein (HDL) [7]. The 'Vroman' effect is not only observed as a maximum in the amount of adsorbed protein as a function of time but also as a maximum in the amount of adsorbed protein as a function of the plasma dilution [8]. Brash et al. [9] observed a maximum in the amount of Fg adsorbed onto glass, silanized glass and polyethylene when they measured the adsorption after 5 min as a function of the plasma
70
"
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
Figure 1. Repeating unit of poly(n-alkylmethacrylate).
dilution. At low plasma concentrations the Fg concentration is also low and therefore the adsorbed amount of Fg will be low. At a high plasma concentration, `he concentrations of the proteins are relatively high, leading to a fast exchange of adsorbed Fg molecules with other protein molecules in solution, for example HMWK. At intermediate plasma concentrations a relatively large amount of Fg may adsorb which will be slowly displaced by other proteins or will be irreversibly adsorbed. Because the 'Vroman' effect involves exchanges of various adsorbed proteins with proteins from a protein solution or plasma, it is expected that this effect is highly dependent on the properties of the material surface. We have previously reported on the surface mobility and structural transitions of (PAMAs Fig. 1) measured by dynamic contact angle poly(n-alkyl methacrylates) measurements [ 10] . The receding angle of PAMAs as a function of side chain length (n) measured at room temperature is gradually increasing from n = 1 to 10. At longer chain lengths (n = 12-18) a sharp decrease in the receding angle is observed indicating that reorientation of side chains takes place which results in relatively hydrophilic surfaces in an aqueous environment. For the adsorption studies of proteins adsorbing from plasma we have selected PAMAs(n = 1, 8 and 18). PAMA(n = 1) is the most hydrophilic polymer of the three PAMAs because it has a small hydrophobic n-alkyl group compared to = = a higher advancing and 8 and exhibited both 8) PAMAs(n 18). PAMA(n = to the relatively long contact to due PAMA(n 1) receding angle compared = For tail of the ester hydrophobic PAMA(n 18) the highest advancing group. contact angle of the three PAMAs was observed due to the long hydrophobic octadecyl group. However, the receding contact angle observed for PAMA(n = 18) was lower than the receding angles observed for the two other PAMAs. The low receding contact angle measured for PAMA(n = 18) could be explained by a reorientation of surface groups and polymer segments after contact of the polymer surface with water. A similar reorientation is unlikely in the case of PAMAs(n = 1 and 8) because the surface groups and segments of these polymers are in a rigid state while those of PAMA(n = 18) are in a mobile state. The aim of this work was to investigate the adsorption and exchange behaviour of proteins adsorbing from plasma onto PAMAs(n = 1, 8 and 18) as a function of time and plasma concentration (Vroman effect) and to relate the results to the surface characteristics of the different PAMAs used. The proteins studied were HSA, Fg, IgG, HMWK and HDL. A two step enzyme-immunoassay (EIA) was used to of the proteins adsorbed to the different measure relative surface concentrations surfaces.
71 MATERIALSAND METHODS
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
Polymer surfaces PAMAs(n = 1, 8 and 18), were synthesized by radical Poly(n-alkyl methacrylates), polymerization of the corresponding n-alkyl methacrylates as described before (10). Polymer surfaces were obtained by dip-coating clean glass coverslips (1.8 x 1.8 cm, Chance Propper LTD, Swethwick, Warley, UK) in 3 070(w/v) solutions of PAMA in toluene P.A. (Merck, Darmstadt, FRG). The coating was dried overnight in vacuo at 50-60°C. To prevent detachment of the very hydrophobic PAMA(n = 8) from the glass slides during immersion in water, glass slides were pretreated with a solution of trimethylchlorosilane (7 v/vO7o, Merck) and pyridine P.A.(7 v/vO7o, Merck) in toluene at room temperature for 1 h. The silanized glass slides were first rinsed with ethanol P.A. (Merck) and ultrafiltrated water and then dried at 110°C in vacuo for 1 h. The slides were then dip-coated with PAMA(n = 8) as described before. Characterization
of surfaces
Polymer surfaces were characterized by dynamic contact angle measurements using a Wilhelmy plate technique as described in ref. [10]. Water contact angles of the polymer surfaces were measured at 20°C. Protein
adsorption
from plasma
In order to measure the adsorption of proteins from diluted human plasma to polymer coated glass slides, a two step enzyme-immunoassay (EIA) was used of al. Breemhaar et according to the method [7] and modified by Poot [11]. The principle of the method is illustrated in Fig. 2. The coated glass slides were mounted in a special test device. The test device enables 24 adsorption experiments with two polymer coated glass slides. The surface area of the polymer coating in each well of the test device is 0.9 cm2. The total volume of each well is approximately 800,ul. In order to prevent the formation of an air-liquid-solid interface upon contact of with 24 the coated the wells were filled with 200,u1 of plasma (solution) glass, buffered saline Darmstadt, (PBS, NaCl(Merck, FRG) 0.140 phosphate ImmediNa2HP04(Merck) NaH2po4 (Merck) 1.3.10-3 mol l-1). thereafter the was started of adsorption experiment ately by adding 200 ,ul (diluted)
Figure 2. Principle of the enzyme-immunoassay(EIA) for detection of adsorbed proteins.
72
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
human
plasma (single donor, Bloedbank, Twente-Achterhoek, Enschede, The Netherlands). During pipetting of the plasma (solution), the end of the pipettor tip did not touch the test surface and was kept under the liquid surface. After gently mixing using the pipettor, the wells were covered with tape to avoid evaporation of water. After the desired adsorption time the wells were rinsed several times with PBS containing 0.005OJo(w/v) Tween-20 (Sigma, St. Louis, USA). The first step of the EIA was started by adding 200,ul of the first antibody solution (1%(v/v)). In the case of HSA, Fg and IgG the first antibodies were produced by rabbits and obtained from the Central Laboratory of the Netherlands Red Cross Blood Transfusion Service (CLB, Amsterdam, The Netherlands). In the case of HDL the first antibody was also a rabbit antibody (Behringwerke AG, Marburg, FRG). The first antibody directed against HMWK was a goat antibody (Dept. of Haematology, University After an exposure time of 1 h the first antiHospital, Utrecht, The Netherlands). body solution was removed and the wells were rinsed with PBS containing Tween-20. Then 200//1 of a solution of the enzyme-labeled second 0.005%(w/v) antibody (conjugate) (United States Biochemical Co, Cleveland, USA) was added. This second antibody was either sheep anti-rabbit IgG (200 times diluted) or rabbit anti-goat IgG (6000 times diluted) depending on the type of first antibody. All first and second antibodies were polyclonal. The bound enzyme was horse-radish peroxidase. After 1 h the conjugate solution was removed. After rinsing with PBS-Tween20 solution, 200,u1 of a solution of the leuko dye (0.1 g/1, 3,3,5,5-tetramethylbenzidine) ((Fluka AG, Buchs, Switzerland) and the enzyme substrate (3olo(v/v) hydrogen peroxide (Merck)) was pipetted into the wells. The reaction was terminated after 30 min by addition of 100 ,ul 2 M sulfuric acid (Merck) and the absorbance of the dye solution was measured at 450 nm (Reader Micro Elisa System, Organon Teknika, Turnhout, Belgium). In order to minimize possible errors in the results of experiments in which protein adsorption was measured as a function of time, these experiments were started at different moments in order to add the first antibody solution into all wells at the same moment. RESULTS Polymer
surfaces
The results of the characterization of the polymer surfaces by dynamic contact angle measurements are summarized in Table 1. The advancing contact angles are higher for PAMAs with longer n-alkyl side chains. After contact with water, surface groups and segments of PAMA(n = 18) are able to reorient. The reorientation results in a lower receding contact angle in the case of PAMA(n = 18) (48°) as compared to the receding contact angle observed for PAMAs(n = 1 (65°) and 8 (73°)). Protein
adsorption
from plasma
Because the results of the EIA experiments showed a day to day fluctuation, in each experiment two materials were compared using PAMA(n = 1) as a reference material. All absorbance values given in the figures are the averages of data obtained
73
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
Table 1. Water contact angles of PAMA surfaces at T = 20°C.
s.d.: standard deviation from four independent experiments, which were carried out simultaneously. Absorbance values presented in the figures were considered to be significantly different when these values differed more than the sum of their standard deviations. Adsorption as a function of time. In experiments in which protein adsorption was measured as a function of time (15 s-20 h), two plasma dilutions (1 :1and 1 : 1000) were used. The adsorption of the following proteins was studied: albumin (HSA), G (IgG), high density lipoprotein (HDL) and high fibrinogen (Fg), immunoglobulin molecular weight kininogen (HMWK).
Figure 3. Adsorption kinetics of proteins adsorbed from 1:1diluted plasma onto PAMAs(n = 1 and 8). Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) IgG adsorption; (d) HDL adsorption.
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
74
Figure4. Adsorption kineticsof proteins adsorbed from 1:1diluted plasma onto PAMAS(n = 1 and 18). Each absorbance value (A450)is the average value obtained from four experiments.Bars indicate standard deviations. (a) HSA adsorption; (b) Fb adsorption; (c) IgG adsorption; (d) HDL adsorption. In Fig. 3 the amounts of protein adsorbed from 1:11 diluted plasma onto PAMAs(n = 1 and 8) surfaces are compared. For HSA and HDL a significantly higher adsorption onto PAMA(n = 8) as compared to PAMA(n = 1) was found for all exposure times. At long exposure times ( - 12 h) the IgG surface concentration on PAMA(n = 1) equals the one on PAMA(n = 8), while the surface concentration of = is 17 min. on after t = Fg only higher PAMA(n 8) A significant increase of the amount of HDL as a function of time is found for on PAMA(n = 1) an increase of the amount of IgG PAMA(n = 8). Furthermore, at long exposure times is observed. In Fig. 4 the results of experiments with regard to the adsorption of proteins from 1:1diluted plasma onto PAMAs(n = 1 and 18) are shown. Differences between the amounts of adsorbed protein to PAMA(n = 1) and onto PAMA(n = 18) were only significant in the case of IgG. The amounts of adsorbed IgG were always larger in the case of PAMA(n = 18) as compared to PAMA(n = 1 ) except for the longest adsorption time where no significant difference was observed. After longer adsorption times the amounts of HSA and HDL adsorbed to both polymers increased 1 and 18) as a slightly. No change in the amounts of Fg adsorbed to PAMAs(n = function of time could be observed. In Fig. 5 the adsorption of HSA, Fg, IgG and HDL from 1:1000 diluted plasma onto PAMAs(n = 1 and 8) is shown. The amounts of HSA and Fg adsorbed onto PAMA(n = 1 ) increased as a function of time, while no significant change in the amounts of adsorbed HSA and Fg to PAMA(n = 8) could be observed. No change in the amount of IgG adsorbed to PAMA(n = 8) was found, while the amount
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
75
diluted plasma onto PAMAs(n = 1 and Figure 5. Adsorption kineticsof proteins adsorbed from 1 : 1000 8). Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) IgG adsorption; (d) HDL adsorption.
Figure 6. Adsorption kineticsof proteins adsorbed from 1:1000diluted plasma onto PAMAs(n = 1 and 18). Each absorbance value (A450)is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) IgG adsorption; (d) HDL adsorption.
76
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
adsorbed to PAMA(n = 1) varied as a function of time. The amount of adsorbed HDL to PAMA(n = 8) increased significantly as a function of time and reached a high value compared to the values observed in the case of PAMA(n = 1). In Fig. 6 the amounts of adsorbed HSA, Fg, IgG and HDL from 1 : 1000 diluted plasma to PAMAs(n = 1 and 18) are compared. Only the differences between the amounts of adsorbed Fg to these surfaces were significant. The amount of Fg adsorbed to PAMA(n = 1) was lower at short exposure times and increased somewhat faster than the amount of Fg adsorbed to PAMA(n = 18). For both PAMAs a minimum in the adsorbed amount of IgG as a function of the adsorption time was found. Adsorption as a function of plasma dilution. The adsorption time was 1 h for experiments in which protein adsorption was measured as a function of the plasma dilution (1:1 to 1:100000). In a series of experiments carried out in 1 day the amounts of protein adsorbed from plasma solutions to two different PAMA surfaces were measured. In Fig. 7 the adsorption of HSA, Fg, HMWK and HDL from plasma solutions is given for PAMAs(n = 1 and 8). Significant differences between the amounts of HSA and HDL adsorbed to PAMAs(n = 1 and 8) were observed. In both cases the amounts of adsorbed protein to PAMA(n = 8) were higher than those adsorbed to At intermediate plasma dilutions the PAMA(n = 1) at high plasma concentrations. amount of Fg adsorbed to PAMA(n = 1) was slightly higher than the amount
Figure 7. Protein adsorption from plasma to PAMAs(n = 1 and 8) as a function of the plasma dilution. The adsorption time was 1 h. Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) HMWK adsorption; (d) HDL adsorption.
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
77
Figure8. Protein adsorption from plasma to PAMAs(n = 1 and 18)as a function of the plasma dilution. The adsorption time was 1 h. Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) HMWK adsorption; (d) HDL adsorption. adsorbed to PAMA(n = 8). For both Fg and HMWK a maximum in the amount of adsorbed protein was found at intermediate plasma dilution. In Fig. 8 the results of similar adsorption are shown for experiments = 1 and Almost no were differences observed between PAMAs(n 18). significant the amounts of protein adsorbed to PAMAs(n = 1 and 18). For HSA, Fg and HMWK a maximum in the adsorbed amount of protein was found at intermediate plasma dilution (1:10,000 to 1:1000). In the case of HDL an adsorption plateau was present at high plasma concentrations. In Fig. 9 protein adsorption of PAMA(n = 8) is compared with the adsorption to PAMA(n = 18). Adsorption of HSA and HDL to PAMA(n = 8) was significantly to higher compared with PAMA(n = 18). On the other hand Fg adsorption = was in Also these maxima in the amounts of PAMA(n 18) higher. experiments adsorbed Fg and HMWK were found at intermediate plasma dilution. DISCUSSION In spite of numerous studies, the exact composition of the protein layer adsorbed to a polymer surface from a multicomponent mixture such as plasma is not yet known. The reason for this is that it is extremely difficult to quantitatively detect one specific protein, often present in low surface concentration, among many others which are adsorbed. Promising techniques to study protein adsorption from complex mixtures are In this investigation a two step enzyme-immunoassay enzyme-immunoassays.
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
78
Figure 9. Protein adsorption from plasma to PAMAs(n = 8 and 18)as a function of the plasma dilution. The adsorption time was 1 h. Each absorbance value (A450) is the average value obtained from four experiments. Bars indicate standard deviations. (a) HSA adsorption; (b) Fg adsorption; (c) HMWK adsorption; (d) HDL adsorption.
[7, 11, 24] has been used to detect the adsorption of HSA, IgG, Fg, HDL and HMWK from human plasma to PAMA surfaces. The advantage of a two step is that usually only one enzyme-labeled enzyme-immunoassay antibody is used, whereas in the case of a one step assay many enzyme-labeled antibodies are necessary. The relative adsorption data are expressed as absorbances of the generated dye in the enzyme-immunoassays. These data are not directly related to amounts of adsorbed Protein molecules conformational protein/cm2. generally undergo and after As a several changes during adsorption. consequence antigenic determinants of a protein molecule may lose their specific structure and may not be able to react with binding sites of the applied antibody. This phenomenon not only material on the and the surface but also on the depends protein possibility of the the is time unfold and which to surface, protein dependent and spread upon influenced by the surface concentration of proteins already adsorbed. Moreover, in the case of multiple layer adsorption, sublayers cannot be detected by the enzymeimmunoassays. Therefore a comparison of the data obtained by this technique with those of a quantitative method like radiolabeling is very difficult. When polymer surfaces are exposed to (diluted) plasma, the composition of the diffusion rates of initially adsorbed protein layer depends on protein concentrations, the proteins and free energies of adsorption. It is expected that plasma proteins which are present in a high concentration and which are small adsorb first. In course of time an exchange of the adsorbed proteins with more strongly adsorbing proteins
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
79 may take place (Vroman effect). Vroman and Adams [12] found that proteins which adsorb from plasma onto glass, displace each other in the following sequence: albumin, IgG, fibrinogen, fibronectin, HMWK and factor XII; i.e. from the most concentrated to the least concentrated protein. In the experiments in which protein adsorption from 1:1 diluted plasma to the PAMAs was studied as a function of time, no significant maxima in the amounts of adsorbed HSA, Fg, IgG and HDL were detected (Figs 3 and 4). The shortest time of exposure of the PAMAs to plasma that could be realized with the EIA technique was 15 s. Probably, an exchange of the adsorbed proteins with proteins in plasma had already taken place within 15 s and had resulted in an adsorbed protein layer in which a displacement of proteins was (practically) impossible. Only the amount of HDL adsorbed to PAMA(n = 8) increased significantly after 20 min of exposure. Probably HDL adsorbs irreversibly to PAMA(n = 8), and not to PAMAs(n = 1 and 18) because PAMA(n = 8) is more hydrophobic than the other two (Table 1). For both series it was found that after long exposure times the adsorption of IgG onto PAMA(n = 1) increased considerably. This phenomenon cannot be explained at the moment. An important result from these experiments is the high adsorption of all proteins 1 and 18). A onto PAMA(n = 8) compared with the adsorption onto PAMAs(n = between and has interaction surfaces been observed by proteins strong hydrophobic several investigators [13-18]. In the case of a hydrophobic surface water molecules are ordered in an ice-like structure at the surface and have a much lower entropy than water molecules in the bulk. The interaction between a hydrophobic surface and a protein originates mainly from an entropy gain due to water desorption from the solid surface and from the protein molecule [ 19-22] . The low HSA adsorption from 1:1 diluted plasma to PAMA(n = 1 ) and PAMA(n = 18) as compared with the adsorption to PAMA(n = 8) may be due to the fact that HSA is more reversibly adsorbed to PAMAs(n = 1 and 18) compared with PAMA(n = 8) which has the most hydrophobic surface of the three polymers. The results of protein adsorption from 1:1000 diluted plasma (Figs 5 and 6) to the PAMAs as a function of time are roughly comparable with those obtained with 1 :1 diluted plasma. The most striking difference between the adsorption values obtained with the two plasma dilutions is the high adsorption of HSA onto PAMA(n = 8) with regard to PAMA(n = 1) at a plasma dilution of 1 : (Fig. 3) compared to the HSA adsorption from 1:1000 diluted plasma (Fig. 5). A high Fg adsorption from 1:1000 diluted plasma was observed for all three PAMAs. The high adsorption of Fg from 1:1000 diluted plasma to the PAMAs may be caused by a lower surface concentration of plasma proteins in the beginning of the adsorption process as compared to PAMAs which are exposed to 1:1diluted Hence plasma. Fg molecules have the opportunity to spread out or unfold, resulting in irreversible protein adsorption [23]. As a consequence, adsorbed Fg molecules will hardly be exchanged with proteins present in the plasma solution. In the case of 1 : diluted plasma, the surface will be totally occupied by proteins, and molecules of Fg which are adsorbed, have a minor opportunity to unfold or spread out upon the surface. Such molecules may be exchanged with protein molecules from the plasma solution. It has been shown by Breemhaar et al. [7]. Poot [11]] and Beugeling [23] that HDL is most probably involved in the displacement of proteins adsorbed from plasma to
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
80 hydrophobic polymers like PVC, polystyrene and polyethylene. It is likely that HDL is also involved in the displacement of proteins, adsorbed from (diluted) plasma to the PAMA surfaces. The relatively large amounts of HDL which adsorb from 1 :1 and 1:1000 diluted plasma to PAMA(n = 8) (Figs 3-6) may support this assumption. However, Figs 4 and 6 show rather low adsorption values for HDL on PAMA(n = 1) and PAMA(n = 18). These low values may be due to the fact that only parts of the HDL molecule consist of apoprotein A-1 to which the first antiof antibody body is directed. This results in a relatively low surface concentration which in turn leads to a low absorbance of the molecules, generated dye (A450 value). The minima in the amount of IgG adsorbed from 1:1000 diluted plasma to 1 and 18) as a function of time could not be explained. PAMAs(n = The 'Vroman effect' is not only observed as a maximum in the adsorption of proteins from plasma as a function of time but also as a maximum in the adsorption of time [9]. proteins as a function of the plasma dilution at a fixed adsorption In general, we also found maxima in the amounts of adsorbed protein as a function of plasma dilution. This was observed for HSA adsorbed to PAMAs(n = 1 and 1, 8 and 18) (Figs 7b, 18) (Figs 7a, 8a and 9a), and for Fg adsorbed to PAMAs(n = 8b and 9b) as well as for HMWK adsorbed to PAMAs(n = 1, 8 and 18) (Figs 7c, 8c and 9c). At high plasma dilution the amounts of proteins adsorbed onto a material surface are low. At high plasma concentrations the amount of a particular protein adsorbed to the surface may also be low because the adsorbed protein may have been displaced by other proteins. In several studies in which glass or glass-like surfaces and polymers were exposed to plasma it was demonstrated that the amount of adsorbed Fg was larger if HMWK-deficient was used [2, 6, 11, or blood plasma 24-26]. This indicates that HMWK may play an important role in the exchange of Fg adsorbed to material surfaces. It is expected that surface hydrophilicity/ role in the exchange of adsorbed protein hydrophobicity plays an important molecules with proteins in solution. When proteins are adsorbed more strongly the exchange is expected to be slower. Vroman et al. [2] contacted surfaces with different hydrophilicity/hydrophobicity with human plasma. They observed platelet adhesion onto the hydrophobic surfaces and no platelet adhesion to glass-like surfaces after adsorption of proteins from plasma. Because platelets preferentially adhere to fibrinogen-coated surfaces they concluded that on the glass-like materials, Fg was exchanged with other proteins while on the hydrophobic surface Fg was not exchanged. Wojciechowski et al. [27] studied the transient adsorption of fibrinogen (Vroman effect) using glass, silylated glass and polymer surfaces. None of the measured adsorption could be correlated with the water contact characteristics angles of the different surfaces. Elwing et al. [28] studied the competition between Fg and HMWK during adsorption on different surfaces after incubation with human plasma. These authors used silanized silicon wafers with a wettability gradient over the surface. It was observed that for all locations at the surface a decrease in anti-fibrinogen binding as a function of time was not associated with an increased anti-HMWK binding. This may indicate that other proteins than HMWK are also involved in the Fg displacement. Mulzer and Brash [29] found that protein layers formed on However, cellulose hemodialysis membranes of poly (methyl methacrylate), Cuprophane,
81
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
acetate and saponified cellulose ester during clinical use contained significant amounts of HMWK, while small amounts of adsorbed Fg seemed to be present. It was suggested that Fg adsorbed from the blood to the membranes had been displaced by HMWK. In this respect it must be mentioned that the authors did not investigate the adsorption of HDL to the dialysis membranes. The adsorption studies with respect to the three PAMA surfaces show that the This amounts of adsorbed Fg and HMWK are low at high plasma concentrations. indicates that both Fg and HMWK are exchanged with (a) more strongly adsorbing protein(s). It has been mentioned above that HDL is most probably involved in this exchange process. The course of protein adsorption to the PAMA surfaces as a function of the plasma dilution (Figs 7-9) will now be explained in more detail. At a high plasma dilution the adsorbed amount of protein is either low due to depletion of the solution or due to diffusion limitation. In order to get a rough estimate of the effects of diffusion and depletion on protein adsorption in our experiments, a simplified model has been used. In this model the effects of convection are neglected, although during the initial few seconds plasma (solution) had to be mixed with buffer in the wells. Furthermore it was intended to calculate maximal amounts of adsorbed protein as a function of time, neglecting the fact that all collissions of protein molecules with the surface may not lead to adsorption. The flux of a protein towards a surface can be calculated using Eqn (1).
in which JX - o is the flux of a protein towards a water/polymer interface assuming diffusion of the protein from a semi-infinite medium towards a flat plate. When every protein molecule which collides with the surface adsorbs, Jx == 0 represents the = the concentration maximal adsorption rate at static conditions at c(t = 0, x > 0), c(t 2:: 0, x = 0) = 0 and D is the diffusion coefficient of the protein. The maximal amount of adsorbed protein (rmaX) after a certain time can be calculated from Eqn (2), obtained by integrating Eqn (1).
A reason for the low adsorption values at low plasma concentration may be which is necessary to diffusion limitation. The minimal protein concentration form a monolayer of adsorbed protein in 1 h will be calculated using Eqn (2). Albumin is present in the highest concentration (c = 40 g/1) and has a relatively high diffusion coefficient (D = 6.1 10-' cm2 s-'). The minimal concentration (c) to form a monolayer of side-on HSA molecules (HSA side-on: 0.2 f.1g/cm2) can now be calculated using Eqn (2). It is again assumed that every molecule which collides with the surface adsorbs. This results in c = 3.8 10-3 g/1 or a plasma dilution of 1:10,000. Another possible reason for the low adsorption values could be depletion of the solution which can be expressed in terms of depletion depth. The depletion depth
82
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
(Xp) is defined as the depth where the concentration of the protein is just below 2007o of its original value and can be calculated with Eqn (3) [30].
In the experiments in which protein adsorption was measured as a function of the plasma dilution the adsorption time was always 1 h. Then the depletion depth for a protein with a diffusion coefficient of 4.10-' cm2/s is 0.7 mm independent of the The depletion depth is larger for proteins with higher diffuprotein concentration. sion coefficients. Because the fluid layer in the EIA experiments had a thickness of about 4.5 mm, depletion cannot be the cause of the small amounts of adsorbed proteins at high plasma dilution. From these results it may be concluded that the low amounts of adsorbed protein observed when the plasma is diluted more than 10000 times, are due to diffusion limitation. At such high dilutions the surface is not totally occupied by adsorbed proteins within the first hour of contact but the solution is not depleted. When plasma is diluted less than 10000 times the composition of the adsorbed protein layer after one hour will not only depend on the diffusion rates of the various proteins but also on the competition between these proteins. The maximum in the amount of adsorbed protein observed in the experiments mentioned above is often present at a plasma dilution of about 1:1000. Assuming again that all protein molecules which collide with the surface adsorb, the maximal amount of adsorbed protein on the surface after a particular time can be estimated for each plasma concentration. The results are given in Table 2. The actually adsorbed amounts of protein are expected to be lower than the calculated ones because a collision of a protein molecule with the surface only results in adsorption if the activation energy for adsorption is low and if the interaction forces are sufficiently high [31]. Although maxima in the amounts of adsorbed protein were often observed for adsorption from plasma which was diluted 1000-10000 times, it may be concluded from Table 2 that the adsorbed amounts may still be very low because of the slow diffusion of proteins to the polymer surface. For example in Fig. 7b the amount of adsorbed Fg was maximal for a plasma dilution of 1 : 10000. Based on the model the maximal amount of Fg transported as a result of diffusion towards the surface is only 9.1 ng/cm2. For the other plasma dilutions lower amounts of adsorbed Fg were measured. This suggests that the amounts of Fg adsorbed from diluted or undiluted plasma do not exceed 9.1 ng/cm2 for an adsorption time of one hour. For HMWK this value is even lower: 0.15 ng/cm2. Table 2. Maximal amount of protein adsorbed onto a surface, after 1 h at various plasma dilutions
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
83 Because HSA is a small and relatively hydrophilic protein it was expected that this protein should be easily displaced from the PAMA surfaces by other proteins. This seems to be the case for PAMAs(n = 1 and 18) where low adsorption values for HSA are found at high plasma concentrations (Figs 7a, 8a and 9a). However, plasma solutions to relatively large amounts of HSA adsorb from concentrated PAMA(n = 8) (Figs 7a and 9a). This is also observed in Fig. 4a, which shows the adsorption of HSA from 1 :1diluted plasma to PAMA(n = 8) as a function of time. was earlier explained by assuming that HSA is irreversibly This phenomenon adsorbed to PAMA(n = 8), which has the most hydrophobic surface of the three polymers after contact with an aqueous solution. PAMA(n = 18) with long hydrophobic n-alkyl chains could give rise to strong protein adsorption, but in contrast to PAMAs(n = 1 and 8) surface groups and segments of PAMA(n = 18) can after contact with reorganize in an aqueous solution. Due to this reorganization water a low receding contact angle was observed for PAMA(n = 18) as shown in of the surface groups and Table 1. It may be concluded that the reorganization segments of PAMA(n = 18) after contact with (diluted) plasma results in surface with respect to protein which are comparable to adsorption properties PAMA(n
= 1).
CONCLUSIONS The results of the adsorption experiments in which protein adsorption from plasma to the PAMA surfaces was studied as a function of the plasma dilution suggest diffusion limitation at high plasma dilution. At high plasma concentrations the amounts of HSA, Fg and HMWK adsorbed to the PAMA's are low compared to the amounts adsorbed at intermediate plasma dilution, except for HSA adsorbed to PAMA(n = 8), indicating that these proteins are displaced from the surface by other plasma proteins. In the case of PAMA(n = 8) one of these proteins may be HDL, because at high plasma concentrations higher surface concentrations of HDL adsorbed on PAMA(n = 8) as compared to PAMAs(n = 1 and 18) were observed. Also the experiments, in which protein adsorption from plasma was studied as a function of the adsorption time, show that a strong interaction of HDL with the PAMA(n = 8) surface occurs. Moreover larger amounts of the other proteins adsorbed from plasma solutions to PAMA(n = 8), as compared to PAMAs(n = 1 and 18). There is evidence that the interaction between PAMA(n = 8) and proteins is caused by a strong hydrophobic interaction. Due to reorientation of surface groups and segments of PAMA(n = 18) in an aqueous solution a similar adsorption behaviour of proteins adsorbing from plasma onto PAMAs(n = 18 and 1) is observed. When the adsorption of proteins from plasma was measured as a function of time no decrease in the amounts of adsorbed protein could be observed. This suggests that after the first 15 s of exposure of PAMA surfaces to plasma no displacement of adsorbed proteins takes place. REFERENCES 1. 2. 3. 4.
L. Vroman and A. L. Adams, Surface Sci. 16, 438 (1969). L. Vroman, A. L. Adams, G. C. Fischer, and P. C. Munoz, Blood 55, 156 (1980). L. Vroman, A. L. Adams, J. Colloid Interface Sci. 111, 391 (1986). S. Winters, R. M. Gendreau, R. I. Leininger, and R. J. Jakobsen, Appl. Spectrosc. 36, 404 (1982).
Downloaded by [University of California Santa Barbara] at 23:16 19 June 2016
84 5. R. M. Gendreau, R. I. Leininger, S. Winters and R. J. Jakobsen, in: Biomaterials: Interfacial Phenomena and Applications, Advances in chemistry series 199, p. 371, S. L. Cooper and N. A. Peppas (Eds). A.C.S., Washington D.C (1982). 6. J. L. Brash and P. ten Hove, J. Biomed. Mater. Res. 23, 157 (1989). 7. W. Breemhaar, E. Brinkman, D. J. Ellens, T. Beugelingand A. Bantjes, Biomaterials5, 269 (1984). 8. T. A. Horbett, Thromb Haemostasis 51, 174 (1984). 9. J. L. Brash and P. ten Hove, Thromb Haemostasis 51, 326 (1984). 10. H. S. van Damme, A. H. Hogt and J. Feijen, J. Colloid Interface Sci. 114, 167 (1986). 11. A. Poot, Thesis, Universityof Twente (1989). 12. I. Vroman and A. L. Adams, in: Proteins at Interfaces. p. 154,J. L. Brash and T. A. Horbett, (Eds). ACS SymposiumSeries 343, A.C.S., Washington D.C. (1987). 13. J. L. Brash and V. J. Davidson, ThrombosisRes. 9, 249 (1976). 14. E. Brynda, N. A. Cepalova and M. Stol, J. Colloid Interface Sci. 18, 685 (1984). 15. H. Y. K. Chuang, W. F. King and R. G. Mason, J. Lab. Clin. Medicine 92, 483 (1978). 16. Y. L. Cheng, S. A. Darst, and C. R. Robertson, J. Colloid Interface Sci. 118, 212 (1987). 17. D. E. Dong, J. D. Andrade and D. L. Coleman, J. Biomed. Mater. Res. 21, 683 (1987). 18. F. Grinnell and M. K. Feld, J. Biomed. Mater. Res. 15, 363 (1981). 19. W. J. Dillman Jr., and I. F. Miller, J. Colloid Interface Sci. 44, 221 (1973). 20. P. G. Koutsoukos, W. Norde and J. Lyklema, J. Colloid Interface Sci, 95, 385 (1983). 21. W. Norde and J. Lyklema, J. Colloid Interface Sci. 66, 295 (1978). 22. W. Norde and J. Lyklema, J. Colloid Interface Sci. 71, 350 (1979). 23. T. Beugeling,in: Modern Aspects of Protein Adsorption. Y. F. Missilis(Ed.). In press. 24. J. L. Brash, C. F. Scott, P. ten Hove, P. Wojciechowskiand R. W. Colman, Blood 71, 932 (1988). 25. A. H. Schmaier, L. Silver, A. L. Adams, G. C. Fischer, P. C. Munoz, L. Vroman and R. W. Colman, ThrombosisRes. 33, 51 (1983). 26. A. Poot, T. Beugeling,W. G. van Aken and A. Bantjes, J. Biomed. Mater. Res. 24, 1021(1990). 27. P. Wojciechowski,P. ten Hove and J. L. Brash, J. Colloid Interface Sci. 111, 455(1986). 28. H. Elwing, A. Askendal and I. Lundstrom, J. Biomed. Mater. Res. 21, 1023(1987). 29. S. R. Mulzer and J. L. Brash, J. Biomed. Mater. Res. 23, 1483(1989). 30. W. J. Beek and K. M. K. Muttzall, Transport Phenomena. John Wiley & Sons Ltd., New York (1977). 31. J. D. Andrade and V. Hlady, Ann. NY Acad. Sci. 516, 158(1987).
