Interactions of proteins in human plasma with modified polystyrene resins C. Boisson-Vidal* and J. Jozefonvicz Laboratoire de Recherches sur les Macromolicules, C.S.P., C.N.R.S. D0502, Avenue f. B. Climent, 93430 Villetaneuse, France J. L. Brash Department of Chemical Engineering, McMaster University, Hamilton, Ontario, Canada
Investigations are reported on the composition of protein layers adsorbed from plasma to various modified polystyrene resins. As well as polystyrene itself, polystyrene bearing sulfonate groups in the benzene rings, and polystyrene sulfonate in which the sulfonate groups were converted to amino acid sulfamide, were investigated. Some of these resins were shown in previous work to have anticoagulant properties. To study the adsorption of proteins from plasma, the resins were exposed to citrate anticoagulated human plasma for 3 h. Adsorbed proteins were then eluted sequentially by 1M Tris buffer and 4% SDS solution, and examined by SDS-PAGE. The gel patterns were similar on all resins except polystyrene. From the MWs of the gel bands, the major protein component appeared to be fibrinogen. Smaller amounts of plasminogen, transferrin, albumin, and IgG were also present. In addition, Ouchterlony immunoassay of the
eluates from one resin gave positive identification of complement C3, fibronectin, IgG, and IgM. Many other minor gel bands remain unidentified. A consistent finding for all resins was the presence of plasmin-type fibrinogen degradation products though the amounts varied with resin type. It is concluded from this (and from experiments showing FDP formation when fibrinogen was adsorbed to the resins, from buffer containing a trace of plasminogen) that the functional groups in these materials promote the adsorption of plasminogen and its activation to a plasmin-like molecule. It appears from the substantial quantities of fibrinogen adsorbed to these materials after 3 h exposure to plasma that the Vroman effect (giving transient adsorption of fibrinogen) is not operative on these materials. It is hypothesized that specific interactions occur between fibrinogen and sulfonate groups.
INTRODUCTION
Insoluble polystyrenes bearing sulfonate and amino acid sulfamide groups of various kinds (see Fig. 1) have been shown to interact specifically with different plasma proteins or serine proteases.',' Most of these resins exhibit heparin-like properties in contact with plasma, i.e., they exhibit a specific in-
To whom correspondence should be addressed. Journal of Biomedical Materials Research, Vol. 25, 67-84 (1991) CCC 0021-9304/91/010067-18$04.00 0 1991 John Wiley & Sons, Inc.
68
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
NH
1
CH-R1
1
COOR2
Figure 1. Structure of the modified polystyrene resins.
teraction with ATIII and thrombin. Resins bearing L-arginine methyl ester groups do not exhibit heparin-like activity, but rather exert a direct action on thr~mbin.~ It has been shown that different parameters have an influence on the protein interactions of such biomaterials, e.g., the type of amino acid, the amino acid content, and the ratio of free sulfonate to amino acid substituted sulfonate. Much of the mechanistic work reported thus far on these resins, has been done in buffered solutions of single proteins. But studies of protein adsorption from plasma generally show results considerably different from solutions of single proteins or relatively simple mixtures consisting of a few protein^.^ Quantities of individual adsorbed proteins are generally less from plasma and binding affinities may also be modified. Also dynamic effects involving displacement of one protein by another have been observed.5s In adsorption from plasma, differences in protein composition between the adsorbed layer and plasma have generally been found. Moreover, the composition of the layer varies from surface to surface and is believed to determine the ultimate response of the blood to contact with the material. Therefore it was of interest to study the protein layer composition, and to identify the proteins adsorbed from plasma onto the family of polystyrene resins described above. In particular, it was of interest to see if differences in protein layer composition existed among the different resins which could be related to their blood interactions. For this purpose, proteins adsorbed after 3 h of contact with plasma, were eluted sequentially by 1 M Tris buffer and 4% SDS. SDS polyacrylamide gel electrophoresis (SDS-PAGE) of eluted proteins was performed and showed a multiplicity of components in the eluates from all resins. Positive identifications were made for one resin by immunodiffusion against specific antibodies. These results were correlated to the presence of different chemical groups and to the amino acid and sulfonate content of the resins.
PROTEINS IN HUMAN PLASMA
69
MATERIALS A N D METHODS
Modified polystyrene resins L-Arginyl methyl ester sulfamide polystyrene (PAOM), L-seryl methyl ester sulfamide polystyrene (PSerOM), and sulfonated polystyrene (PSSO3) were prepared according to a method previously described.' L-Glutamic acid sulfamide polystyrene (PSS02Glu) was kindly provided by Dr. C. Fougnot.' The compositions of the resins are given in Table I. The starting material consists of crosslinked polystyrene beads of diameter 37-74 pm, (Biobeads SX2, Bio-Rad, Richmond, CA, USA) and is used as a control (PS). The resins were equilibrated in isotonic Tris buffer pH 7.4.
Proteins Human fibrinogen was obtained from Kabi (Stockholm, Sweden). The lyophilized product was dissolved in distilled water and dialyzed overnight against an appropriate buffer, usually 0.05 M Tris pH 7.4, and stored frozen until used. This material is known to contain traces of immunoglobulin, a2 macroglobulin, and plasminogen (less than 0.2%). Other suspected contaminants are fibronectin and fibrinogen degradation products. SDS PAGE of this fibrinogen shows, after reduction with /3-mercaptoethanol,in addition to the expected bands for the a, /3, and y chains of fibrinogen (molecular weights (MWs) of 67,56, and 53 kD, respectively)additional bands at higher and lower MW consistent with the above mentioned contaminants (see Fig. 3). Fibrinogen degradation products (FDP) were obtained by treatment of Kabi fibrinogen with streptokinase. After various times, €-amino caproic acid was added to quench the reaction. Human plasminogen was prepared in our laboratory" and kindly provided by S. Khamlichi. Human albumin was from Behringwerke (Marburg, West Germany) and was 100% immunochemically pure. Porcine heparin was obtained from McMaster University Hospital TABLE I Chemical Composition of Polystyrene Resinsa Amino Acid Substituted Benzene Rings ~
Unsubstituted Benzene Rings
Sulfonated Benzene Rings
PSSOS PAOM5
10,o 14,5
90,O 80,5
-
-
-
5,o
(CH213 NH
CH3
PAOM71 PSS02Glu PSerOMlO
19,O 22,5 14,5
10,o
71,O 3,6 10,o
NH,-C=NH (CH&-COOH CHzOH
CH3 H CH3
73,9 75,5
R1
"Data are given as percent of benzene rings in specified form.
I
R2
70
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
Hamilton, Canada. Agkistrodon Rhodostromm (snake venom, Arvin) was kindly provided by Dr. M.W C. Hatton (McMaster University). Sepharose 6B was from Pharmacia, Ltd (Montreal, Canada). Eagle’s Minimum Essential Medium (MEM) was bought from Gibco (Burlington,Ontario, Canada). Antisera to human plasma proteins were from the following sources: anti haptoglobin, anti C3, anti Fc fragment of IgG, anti antithrombin 111, anti a2 macroglobulin, anti fibronectin, anti IgM and anti a1 antitrypsin were from Cappel Cochranville, PA, U.S.A. Anti IgG (H & L chains) and antilipoproteins were from Miles Scientific, Rexdale, Ontario, Canada.
Blood and plasma Human platelet poor plasma (PPP) was prepared from acid-citrate dextrose anticoagulated blood (6 vol blood: 1 vol ACD solution). The whole blood was taken from normal, healthy, single donors into citrate anticoagulant and centrifuged (3 min 12008, 3 min 1200g, 15 min 25008 at 20°C). The plasma was then stored at -70°C. Arvinized plasma was prepared as previously described’ from fresh human blood collected into ACD.
Heparinized plasma The heparin was contained in vials, each containing 100 anticoagulant IU/mL in isotonic saline solution. To allow for easier mixing into the blood, the heparin was diluted with isotonic Tris buffer to 15 mL. It was then mixed with 150 mL of freshly drawn blood (single donor), and processed in the usual way to obtain platelet poor plasma.
Electrophoresis and gels Acrylamide, bisacrylamide, SDS, silver nitrate, Coomassie blue R250 and other electrophoresis products were from Bio-Rad (Richmond, CA, USA). Molecular weight standards were the low and high MWs calibration kits from either Pharmacia (Uppsala, Sweden), Bio-Rad, or Sigma (St. Louis, MO, USA). Sepharose-heparin was prepared using the CNBr method of Miller Andersson et al.” Double diffusion immunoassay: Ouchterlony double diffusion immunoassays were performed on the samples eluted from the PAOM5 resin by 1M Tris buffer. Both antigens and antisera were placed into wells punched out of Agarose gel (l%), and were allowed to diffuse toward each other.” Following diffusion, the gels were stained with 1.5%Naphtol Blue Black B solution.
PROTEINS IN HUMAN PLASMA
71
Adsorption procedures
Four grams of resin particles suspended in isotonic Tris buffer pH 7.4, was packed into a glass column 17.5 cm x 1.3 cm. Ten milliliters of the plasma to be studied was then loaded onto the column. The system was incubated for various times from zero (direct elution) to 3 h at room temperature. The latter is a period sufficiently long that adsorption should have reached "equilibrium." The plasma was then removed by washing with isotonic Tris buffer until the adsorbance at 280 nm of the effluent returned to zero. The column was then eluted sequentially with 1M Tris pH 7.4, and with 4% SDS in isotonic Tris buffer. Eluted samples were concentrated by ultrafiltration (Amicon model 8010, 10 mL) and studied by SDS PAGE using 11% and 5 to 15% gradient gels.I3Samples were reduced by boiling in p-mercaptoethanol. Following electrophoresis, the gels were stained for protein with 0.05% Coomassie brilliant blue solution. Silver staining was used for selected cases.
RESULTS
Exposure of resins to PPP
The characterization of proteins adsorbed from plasma to polystyrene resins was performed after 3 h of contact at room temperature. A typical chromatogram from a column experiment is shown in Figure 2. Following
t1
iISOTONIC
I /I
TRlS
4%
SDS
100
FRACT ION NUMBER
b
Figure 2. Typical chromatogram for elution of ACD plasma from PAOM 5 resin.
72
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
the wash step with isotonic Tris buffer, sharp peaks were eluted both by 1M Tris pH 7.4, and 4% SDS in isotonic Tris for the resins PS, PSSO3, PAOM5, PSS02Glu, and PSerOM10. To check whether significant amounts of protein remained on the column after treatment with SDS, elution with 4M urea was performed. Only a very small peak was generally eluted and contained mostly fibrinogen, FDP, and albumin. Different elution patterns were observed with PAOM71. When ACD plasma was incubated with this resin, it clotted immediately. This behavior was attributed to the high concentration of positive charge present in PAOM71 compared to PAOM5. We hypothesized that when ACD plasma was contacted with PAOM71, the citrate ions formed a complex with the positive charge leaving the calcium free, and thus causing the plasma to coagulate. However, when the resin was pretreated with 2M trisodium citrate (referred to PAOM7lc) in order to saturate the guanidyl groups, no clotting occurred, and sharp peak was eluted as for the other resins. Such a pretreatment, however, presumably masks the effect of the guanidyl groups. Thus in order to study the influence of the free guanidyl groups on protein adsorption, heparinized plasma was contacted with PAOM71 and the results were compared with those obtained in the same conditions with ACD plasma and PAOM7lc. No significant differences were detected (data not shown). Figure 3 shows the SDS PAGE from a typical experiment using normal ACD plasma (3 h contact time) on PAOM5. The protein staining bands were tentatively identified by their position in the gel in relation to molecular weight standards. An experiment using a shorter contact time (no incubation) was undertaken to investigate possible kinetic effects and almost identical results were obtained (data not shown). As can be seen in Figure 3, the 1M Tris peak, lanes 2 and 3 (the latter analyses the end of the 1M Tris peak) shows bands at 94 kD (A, probably plasminogen), 80 kD (B, transferrin), and 69 kD (C, albumin).14In addition, a group of bands centred at about 40 kD (D) and the band at 14 kD (E) may be attributed to plasmin-induced FDP (see lane 6). The SDS peak (lane 4) also shows bands corresponding to plasminogen (A), transferrin (B) and albumin (C). The principal component, however, is fibrinogen (see lane 5) and some is degraded (lane E). The albumin and the aa band of fibrinogen run at the same place on the gel. Figures 4 and 5 show, respectively, the gels of the 1M Tris and SDS eluates from PS (lane 8), PSSO3 (lane 7), PSS02Glu (lane 6), PSerOMlO (lane 5), PAOM5 (lane 4)and PAOM71 (lane 3). Gels of FDP and fibrinogen are shown for comparison (lanes 1 and 2). No proteins were eluted in the 1M Tris peak from the PS column. For the other resins, the 1M Tris peak shows major bands at about 160 kD (A, PAOM71), 123 kD (B, PAOM71 and PSSO3), 118 kD (C, PAOM5 and PSSO3), 110 kD (probably plasminogen (Pl) PAOM71, PAOM5, PSerOM10, and PSSO3), 100 kD (D, PAOM71, PAOM5, PSerOM10, and PSSO3), 80 kD (transferrin (Tr) PAOM71, PAOM5, and PSS02Glu), 67 kD (albumin (Alb), PAOM71, PAOM5, PSerOM10, PSS02Glu, and PSSO3), 67-56 and 53 kD (a,p and y chains of fibrinogen (Fg), PAOM71, PAOM5, and PSS02Glu), 53-49 and 43 kD (FDP, PAOM71, PAOM5, PSerOM10, PSS02Glu, and PSSO3), 28 kD
PROTEINS IN HUMAN PLASMA
73
Figure 3. SDS PAGE of proteins eluted from PAOM after incubation with ACD plasma (3 h contact time). Samples were reduced in 2% p-mercaptoethanol and loaded onto an 11%gel. Lane 1: MW standards. Lanes 2 and 3: 1M Tris eluate. Lane 4: SDS eluate. Lane 5: Kabi fibrinogen (control). Lane 6: fibrinogen degradation products produced by incubation of Kabi fibrinogen with streptokinase for 3 h.
(E) and 20 kD (F). In the SDS peak, the major bands are at 154 kD (A, PAOM5), 108 kD (B, PAOM5), 100 kD (Pl, PAOM5), 81 kD (Tr, PAOM71, PAOM5, PSerOM10, PSS02Glu, and PSSO3), 67,56, and 53 kD (Alb, all resins and Fg, PAOM71, PAOM5, PSerOM10, PSS02Glu, and PSSO3) and groups of band at 30 kD (C) and 20 kD (D) (all resins).14There are also a few weaker bands. The behavior of PAOM7lc, in which guanidyl groups were neutralized by treatment with citrate is shown in Figure 6. In comparison with the same
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
74 1
2
3
4
5
6
7
8
Figure 4. SDS PAGE (5-1572 gradient gel, reducing conditions). Comparison of 1M Tris eluates after incubation of ACD plasma on PS, PSSO3, PSS02Glu, PSerOM10, PAOM5, and heparinized plasma on PAOM71. Lane 1: fibrinogen degradation products produced by incubation of Kabi fibrinogen with streptokinase for 3 h. Lane 2: Kabi fibrinogen (control). Lane 3: PAOM 71. Lane 4: PAOM 5. Lane 5: PSerOM10. Lane 6: PSS02Glu. Lane 7: PSSO3. Lane 8: PS. P1: plasminogen, Tr: transferrin, Alb: albumin, Fg: fibrinogen, FDP: fibrinogen degradation product.
resin in which the free guanidyl groups are available (Figs. 4 and 5, lane 3) there are significant differences. In the 1M Tris eluate, Figure 6 shows less albumin (Alb) and more transferrin (Tr) than Figure 4 (lane 3). In the SDS eluate, Figure 6 shows more fibrinogen (Fg) and less albumin (Alb) than Figure 5 (lane 3). Fibrinogen appears to be the major component of the adsorbed layer on all the resins studied, with the exception of PAOM7lc and PSerOM10. In the case of PAOM5 and PAOM7lc, it is eluted mainly by SDS (Figs. 3 and 6) except for the reconditioned resins, in which case it is eluted by both Tris and SDS (Figs. 4 and 5).
PROTEINS IN HUMAN PLASMA 1
2
75
3
4
5
6
7
8
Figure 5. SDS PAGE (5-15% gradient gel, reducing conditions). Comparison of 4% SDS eluates after incubation of ACD plasma on PS, PSSO3, PSS02Glu, PAOM5, PSerOM10, and heparinized plasma on PAOM71. Lane 1: fibrinogen degradation products produced by incubation of kabi fibrinogen with streptokinase for 3 h. Lane 2: kabi fibrinogen (control). Lane 3: PAOM71. Lane 4:PAOM5. Lane 5: PSerOM10. Lane 6. PSS02Glu. Lane 7: PSSO3. Lane 8: PS. P1: plasminogen, Tr: transferin, Alb: albumin, Fg: fibrinogen, FDP: fibrinogen degradation products.
Part of the eluted fibrinogen is partially degraded as indicated by good correspondence between the three bands 53 kD, 49 kD, and 43 kD in both the 1M Tris and SDS peaks, and corresponding bands in streptokinase induced fibrinogen degradation products (Figs. 4 and 5, lane 1).No FDP are observed on PAOM71 (Fig. 4 lane 3, and Fig. 6).
Exposure of PAOM5 to arvinized plasma Since fibrinogen appeared to dominate the protein layer adsorbed from normal plasma to these resins, it was of interest to investigate adsorption
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
76
1
2
3
4
5
Figure 6. SDS PAGE (ll%,reducing conditions) of proteins eluted from precitrated PAOM71 after contact with ACD plasma (3 h contact time). Lane 1: MW standards. Lane 2: 1M Tris eluate. Lane 3: 4% SDS eluate. Lane 4:Kabi fibrinogen (control). Lane 5: fibrinogen degradation products produced by incubation of Kabi fibrinogen with streptokinase for 3 h.
from plasma which was depleted of fibrinogen. This experiment was performed after 3 h of contact with arvinized plasma from which fibrinogen had been partially removed by treatment with a snake venom clotting enzyme. The concentration of the remaining fibrinogen in this plasma was determined and was found to be about 0.3%of the initial value.’ Figure 7 shows that the main protein components are albumin and high molecular weight proteins, at about 200 kD (A) in the 1M Tris peak and 120 kD (B) in the SDS peak. Many other lower MWs components are also present. Fibrinogen is also detected (lanes Fg) in spite of the low residual concentration. This indicates a high affinity of the protein for the polymer surface.
PROTEINS IN HUMAN PLASMA
77
Figure 7. SDS PAGE (ll%,reducing conditions) of proteins eluted from PAOM5 after contact with Arvinized plasma (3 h contact time). Lane 1: MW standards. Lane 2: 1M Tris eluates. Lane 3: 4% SDS eluates. Lane 4: Kabi fibrinogen. Arrows indicated the 200 kD (in the 1M Tris peak) and 120 kD (in the SDS peak) bands.
To identify the proteins eluted by 1M Tris, Ouchterlony immunodiffusions of column eluates were run against antibodies to various plasma proteins. The results of these analyses are shown in Figure 8. Positive identifications were made for C3 (well l), Fc fragment of IgG (well 3), IgG (well 4), fibronectin (well 7), and IgM (well 8). Such assays were not done for fibrinogen, FDP, albumin, and plasminogen, since it seemed clear from the gels that positive results would have been obtained for these proteins. Tests for other proteins were negative, which may reflect a concentration of the protein below the detection limit of the method. On the basis of the positive results obtained, we can attribute the bands at 200 kD to fibronectin. Human plasma fibronectin is composed of two polypeptides (aand p chains) each with MWs of about 200 kD. SDS PAGE of C3 has shown two polypeptide chains after
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
78
Figure 8. Ouchterlony analysis of 1M Tris fractions eluted from PAOM5 resin after contact with Arvinized plasma. Central wells contains the antisera: 1, anti C3; 2, anti haptoglobin; 3, anti Fc Fragment of IgG; 4,anti IgG; 5, anti d-antitrypsin; 6, anti antithrombin 111; 7, anti fibronectin; 8, anti IgM; 9, anti a2 macroglobulin; 10, anti @lipoprotein. Peripheral wells: 1M Tris eluate diluted to various extents.
denaturation and reduction. The polypeptide chains a and p have been assigned respective MWs of 115 11.5 and 45 k 7 kDs.”
*
Incubation of resins with fibrinogen solution in presence of added plasminogen To provide supporting evidence for the hypothesis that surface-bound fibrinogen in the presence of plasminogen undergoes lysis, experiments were performed with fibrinogen solution (3 mg/mL) to which plasminogen was added (3 kg/mL, i.e., 0.1% (w/w) relative to fibrinogen). Gels of the eluted proteins are shown in Figure 9, which includes controls generated by incubation of fibrinogen with 0.1% (w/w) plasminogen in isotonic Tris buffer. The 1M Tris eluate shows essentially unaltered fibrinogen and plasminogen. The groups of bands near 30 kD probably represents a small amount of FDP. In
PROTEINS IN HUMAN PLASMA
79
Figure 9. SDS PAGE (10-15% gradient gel, reducing conditions) of proteins eluted from resins after contact with fibrinogen solution containing 0.1% plasminogen. (A) Lane 1: low-molecular-weight standard, lanes 2, 3, 4, 5, and 6 : 4% SDS eluates from PSEOM10, PAOM71, PAOM5, PSSO3, and PSSOZGlu, respectively. Fg: fibrinogen, FDP: fibrinogen degradation products. (B) Lane 1: low-molecular-weight standard, lanes 2, 3, 4, 5, and 6 : 1M Tris eluates from PSEOMlO, PAOM71, PAOM5, PSSO3, and PSS02Glu, respectively.
the 4% SDS peak, FDP is present and in an extensively degraded form. In contrast to the other resins, PSerOMlO appears to induce considerably greater and more extensive degradation of fibrinogen. Only a small quantity of intact fibrinogen is present (Fig. 9, lane 5). Table I1 shows data on the relative quantities of fibrinogen plus plasminogen retained and eluted from the different resins. As can be seen, the arginine substituted resins adsorb significantly smaller amounts of the proteins than all the other resins (1M Tris eluate). On all the resins studied the adsorbed proteins are eluted mainly by SDS.
DISCUSSION
The exposure of blood to artificial surfaces leads to a rapid and competitive adsorption of plasma proteins within a matter of s e c ~ n d s .The ~ ” ~composition of the adsorbed layer differs considerably depending on the nature of the material. As can be seen from the gels in Figures 3, 4, and 5, the protein mixtures eluted from the polystyrene resins used in this study are complex and the gels contain many bands. The results of Ouchterlony analysis (Fig. 9) also show the extreme complexity of the process of adsorption from plasma. In general a much smaller fraction of the proteins adsorbed are eluted by 1M Tris (0.5-7%) than by 4% SDS solution (5-18%) (Table 111). This observa-
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
80
tion may be interpreted as showing that most of the proteins are bound by interactions other than electrostatic. The gel patterns of proteins eluted following 3 h of plasma contact show certain similarities on all the resins except the "starting" unmodified polystyrene (PS). Bands corresponding to albumin, plasminogen, fibronectin, IgG, and other unidentified bands (160, 150, and 43 kD) are generally observed. Fibrinogen appears to be a major component in every case. This is perhaps surprising since although fibrinogen is a major protein component of blood and is known to have a high surface affinit~,'~ it has been shown that its adsorption from blood or plasma is transient.>* This is the well known "Vroman effect" by which more abundant, low-surface-affinity proteins are adsorbed initially and later displaced by less abundant, high-surface-affinity proteins, e.g., the contact phase clotting factors." Since there are still substantial quantities of fibrinogen adsorbed to the present resins after 3 h of plasma contact, it appears that the Vroman effect is less pronounced on these materials. In agreement with this conclusion, we have shown recently that the Vroman effect is greatly reduced on polyurethane materials containing sulfonate groups in the hard segments." It appears that specific interactions may occur between sulfonate groups and fibrinogen. Compared to the other resins, the unmodified polystyrene (PS) apparently shows much less adsorption. Perhaps more precisely, less protein is eluted from this resin. This material is strongly hydrophobic with a water contact angle of about 80" and presumably can bind proteins only through hydrophobic interactions. Hydrophobic interactions contribute favorably both to AH and to AS: to the former via van der Waals interactions, and to the latter through dehydration of the interacting surfaces involving a disordering of water molecules. This entropy gain is considered to represent a major driving force for protein adsorption on hydrophobic surfaces. As can be seen from Figures 4 and 5, lane 8, no proteins were eluted by 1M Tris while only a few (some albumin, fibrinogen and probably IgG) were eluted in 4% SDS. These results agree with the idea that electrostatic interactions do not occur on this surface. In addition, they may show that SDS can disrupt van der Waals interactions between proteins and polystyrene. It must be emphasized that proteins are known to be adsorbed to polystyrene but may be particularly strongly bound and difficult to elute esTABLE I1 Relative Quantities (Percent) of Proteins Eluted from Polystyrene Resins by Isotonic Tris, 1M Tris, and 4%SDS Buffers, Following 3 h Contact with Fibrinogen Solution (3 mg/mL) to Which Plasminogen Was Added (3 pg/mL)
PSSO3 PSS02Glu PAOM5 PAOM71 PSerOMlO
Isotonic Tris Buffer
1M Tris
4% SDS
4 10 59 45 4
16.0 25.5 2.0 2.0 29.0
80.0 64.5 38.0 53.0 51.0
81
PROTEINS IN HUMAN PLASMA
TABLE 111 Relative Quantities (Percent) of Proteins Eluted from Reconditioned Polystyrene Resins by Isotonic Tris, 1M Tris, and 4% SDS Buffers, Following 3 h Contact with Plasma ~
PSSO3 PSS02Glu PAOM5 PAOM71 PSerOMlO
Isotonic Tris Buffer
1M Tris
4% SDS
80 89 88 89 75
7.0 5.5 0.5 6.0 7.0
13.0 5.5 13.5 5.0 18.0
pecially after long contact times. We showed previ~usly'~ that IgG, albumin, and fibrinogen adsorb in substantial quantities to polystyrene from both single protein solutions and from plasma. In addition, fibrinogen adsorption from plasma was not transient in contrast to many other surfaces,s8 again suggesting strong binding to polystyrene. More recently, Rapoza and Horbettz0 have shown that the SDS elutability of fibrinogen adsorbed on polystyrene is low compared to other surfaces. It therefore seems more likely that the relative lack of proteins in the eluates from PS is due to difficulty of desorption rather than to low adsorption. The substitution of sulfonate groups in the benzene rings of polystyrene causes a significant increase in hydrophilicity, leading to much greater adsorption of proteins from plasma, or easier elution of adsorbed proteins, notably fibrinogen. On conversion of some of the sulfonate groups to amino acid sulfamide groups, the protein adsorption behavior in plasma does not appear greatly altered. However, there are some important differences among the different amino acid or amino acid ester substituents. Thus, for the highly substituted arginine methyl ester resins, there appears to be less adsorption of proteins elutable in SDS (see Table 111). This result may be related to the relatively small concentration of sulfonate groups in PAOM71 resin. For this resin an experiment was also performed in which the guanidyl groups were "blocked" by treatment with trisodium citrate prior to ACD plasma exposure (see Fig. 6). In the 1M Tris peak, very few proteins are present, and there is much less albumin and virtually no fibrinogen. However a significant amount of plasminogen is observed. The SDS peak contains mostly transferrin, fibrinogen, and possibly some FDP. In general for the PAOM71 resin, whether the guanidyl groups are blocked or free, the generation of FDP is not extensive. It thus seems likely that sulfonate groups play a major role in fibrinogen degradation via plasminogen activation on these resins, as shown by Kichenin-Martin et a1." These authors studied adsorption of plasminogen on PSS02Glu and PSSO3. They have observed that purified plasminogen is adsorbed and becomes enzymatically active on the surface of these modified polystyrenes without any cleavage of the molecule. (Plasminogen is a single-chain protein of MW 94 kD and is normally converted to plasmin by enzymatic cleavage of a peptide bond within a double
82
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
disulphide loop.22This creates an enzymatically active molecule having a light chain of MW 25 kD and a heavy chain of MW 60 kD held together by disulfide bonds. The active site resides on the light chain.) A similar mechanism has been observed for the streptokinase/plasminogeninteraction and involves the formation of a plasminogen-streptokinase complex possessing enzymatic activity similar to that of pla~min.’~ In previous we showed that when glass beads were contacted with purified fibrinogen to which a trace of plasminogen was added, degradation of eluted fibrinogen was very extensive. These results were also interpreted in terms of a surface-mediated activation of plasminogen to plasmin followed by fibrinogenolysi~.~~ In connection with surface “fibrinolytic activity,” taken as the ability of a surface to adsorb and activate plasminogen, the serine methyl ester modified resin (PSerOM10) is of particular interest, and appears to have strong fibrinolytic properties as revealed by the experiments with fibrinogen-plasminogen mixtures (Fig. 9 and Table 11). The mechanism by which this activity is expressed is, for the moment, unknown and may not be the same as for the sulfonate-mediated effect (see above). It is perhaps relevant to point out that the methyl ester of serine is somewhat analogous to the seryl site of serine proteases which converts proenzymes to enzymes, for example, the coagulation factors. Thus, it is possible that plasminogen is converted to plasmin on this surface by the conventional cleavage mechanism.25Further work would be required to investigate this possibility. Most of the plasma data were obtained after 3 h incubation. However, it seemed possible that different adsorption patterns would be seen at different times. Indeed, adsorbed protein layers are known to exchange with proteins in solution and thus the composition of the layer may change over time. Therefore, an experiment was performed on one resin, PAOM5, in which the plasma was passed directly through the column without incubation (data not shown). The adsorption patterns were, in general, similar to those observed at longer contact time. However, it was noted that more intact fibrinogen was present after the short contact time. This finding is in agreement with the idea that, via the Vroman effect, less fibrinogen displacement should have occurred at shorter contact times. Surfaces which retain fibrinogen more strongly, like PSSO3, PSS02Glu and PSerOMlO (Table 111), may inhibit surface-induced coagulation by preventing access of coagulation factors to the surface. But they may, by the same token, be more platelet reactive. In this connection, however, ex vivo animal experiments performed on small-diameter tubings made of polystyrene-polyethylene copolymers grafted with sulfonate and aspartic acid sulfamide groups showed no significant platelet adhesion and aggregation on the surface of the tubings, during the 2 h of the experiment.26This was true whether the tubing was pretreated with either plasma or antithrombin 111. The resins used in this study were shown to interact specifically with thrombin and AT I11 in previous experiments,’ the specific affinity of the two proteins being dependent upon the properties of the resin. From the present studies, however, no clear differences in protein layer composition after
PROTEINS IN HUMAN PLASMA
83
plasma contact existed among the different resins. The slight variations in adsorption from plasma were in terms of relative concentrations rather than type of protein. Sulfonate groups are clearly a factor contributing to the adsorption of the fibrinogen. They may also be involved in adsorption of plasminogen and in the surface-mediated activation of plasminogen to plasmin. Financial support of this work by the Heart and Stroke Foundation of Ontario, the Medical Research Council of Canada, the Natural Sciences and Engineering Research Council of Canada, and the Centre National de la Recherche Scientifique (CNRS UD 0502) of France is gratefully acknowledged.
References 1.
2.
3. 4. 5. 6.
7. 8. 9.
10. 11. 12. 13. 14. 15. 16.
C. Boisson, J. Jozefonvicz, and J.L. Brash, “Adsorption of thrombin from buffer and modified plasma to polystyrene resins containing sulphonate and sulphamide arginyl methyl ester groups,” Biomaterials, 9, 47-52 (1988). C. Fougnot, M. Jozefowicz, and R. D. Rosenberg, ’Adsorption of purified thrombin or antithrombin I11 on two insoluble anticoagulant polystyrene derivatives: I1 Competition with the other plasma proteins,” Biomaterials, 5, 89-93 (1984). C. Boisson, D. Gulino, and J. Jozefonvicz, ’Affinity of antithrombin I11 for insoluble modified polystyrene,” Thromb. Res., 46, 101-108 (1987). J. L. Brash and T. A. Horbett, Proteins at Interfaces, ACS Symposium Series, vol. 343, Amer. Chem. SOC.,Washington D.C., 1987. L. Vroman, A. L. Adams, G.C. Fischer, and P.C. Munoz, ”Interaction of high molecular weight kininogen, factor XI1 and fibrinogen in plasma at interfaces,” Blood, 55, 156-159 (1980). L. Vroman and A. L. Adams, ”Identification of rapid changes at plasmasolid interfaces,” J. Biomed. Muter. Res., 3, 43-67 (1969). T.A. Horbett, “Mass action effects on competitive adsorption of fibrinogen from hemoglobin solutions and from plasma,” Thromb. Haemostas., 51, 174-181 (1984). J. L. Brash and P. ten Hove, “Transient adsorption of fibrinogen on foreign surfaces: similar behavior in plasma and whole blood,” J Biomed. Muter. Res., 23, 157-169 (1989). C. Boisson, D. Gulino, and J. Jozefonvicz, ”PolystyrPnes insolubles modifies par des derives de l’arginine: syntheses, interactions, avec les proteines plasmatiques,” Bull. la Rev. SOC.Chim. France, 4, 84-87 (1985). D.G. Deutsch and E.T. Mertz, ”Plasminogen, purification from human plasma by affinity chromatography,” Science, 170, 1095-1096 (1970). M. Miller-Anderson, H. Borg, and L.O. Anderson, “Purification of ATIII by affinity chromatography,” Thromb. Res., 5, 439-442 (1974). 0. Ouchterlony, Handbook of lmrnunodiffusion and Immuno-electrophoresis,“ Ann Arbor Science Publishers, Ann Arbor, MI, 1968. U. K. Laemmli, ”Cleavage of the structural proteins during assembly of the head of the bacteriophage T4,” Nature, 227, 680-685 (1970). J. L. Brash, S. Uniyal, B. M. C. Chan, and A. Yu, “Fibrinogen glass interactions: a synthesis of recent research,” ACS Symp. Ser., Mars (1983). B.F. Tack and J.W. Prahl, “Third component of human complement: purification from plasm, a and physicochemical characterization,” Biochemistry, 15,4513-4521 (1976). L. Vroman, A. L. Adams, M. Klings, G.C. Fischer, P.C. Munoz, and R. P. Solensky, ”Reactions of formed elements of blood with plasma proteins at interfaces,” Ann. NX Acad. Sci., 283, 65-76 (1977).
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
84
17. S. Uniyal and J.L. Brash, ”Patterns of adsorption of proteins from human plasma onto foreign surfaces,” Thromb. Huemostas., 47, 285-290 (1982). 18. J. L. Brash, C. F. Scott, P. ten Hove, P. Wojciechowski, and R.W. Colman, “Mechanism of the transient adsorption of fibrinogen from plasma to solid surfaces: role of the contact and fibrinolytic systems,” Blood, 71, 932-939 (1988). 19. J. P. Santerre, N. Van Der Kamp, and J. L. Brash, “Effect of sulfonation of segmented polyurethanes on the transient adsorption of fibinogen from plasma: possible correlation with anticoagulant behavior,” Trans. SOC. Biomat., 12, 113 (1989). 20. R.P. Rapoza and T.A. Horbett, “Changes in the SDS elutability of fibrinogen adsorbed from plasma to polymers,” J. Biomater. Sci. Polym. Edn, 1, 99 (1989). 21. V. Kichenin-Martin, L. Pejaudier, and M. Steinbuch, “Unusual, activation mechanism of plasminogen,” Thromb. Res., 52,469-478 (1988). 22. H. R. Lijnen and M. D. Collen, Sem. Thromb. Huemostas., 8, 2-10 (1982). 23. L. A. Schick and F. J. Castellino, “Direct evidence for the formation of an active site in the plasminogen moiety of the streptokinase-human plasminogen activator complex,” Biochem. Biophys. Res. Comm., 57,47-52 (1974). 24. J, L. Brash, B. M.C. Chan, P. Szota, and J. A. Thibodeau, “Degradation of adsorbed fibrinogen by surface generated plasmin,” J. Biomed. Muter. Res., 19, 1017-1029 (1985). 25. E Bachman, ”Fibrinolysis,” in XIth Congress Thrombosis Haemostasis Brussels, 227-265 (1987). 26. V. Migonney, “Fonctionralisation de la surface interne de matCriaux tubulaires. Etude de l’inhibition de la thrombine par l’antithrombine 111?I la surface de ces matCriaux,” ThPse dEtat, Paris XIII, 4 June 1986.
Received January 10, 1990 Accepted July 13, 1990
Investigations are reported on the composition of protein layers adsorbed from plasma to various modified polystyrene resins. As well as polystyrene itself, polystyrene bearing sulfonate groups in the benzene rings, and polystyrene sulfonate in which the sulfonate groups were converted to amino acid sulfamide, were investigated. Some of these resins were shown in previous work to have anticoagulant properties. To study the adsorption of proteins from plasma, the resins were exposed to citrate anticoagulated human plasma for 3 h. Adsorbed proteins were then eluted sequentially by 1M Tris buffer and 4% SDS solution, and examined by SDS-PAGE. The gel patterns were similar on all resins except polystyrene. From the MWs of the gel bands, the major protein component appeared to be fibrinogen. Smaller amounts of plasminogen, transferrin, albumin, and IgG were also present. In addition, Ouchterlony immunoassay of the
eluates from one resin gave positive identification of complement C3, fibronectin, IgG, and IgM. Many other minor gel bands remain unidentified. A consistent finding for all resins was the presence of plasmin-type fibrinogen degradation products though the amounts varied with resin type. It is concluded from this (and from experiments showing FDP formation when fibrinogen was adsorbed to the resins, from buffer containing a trace of plasminogen) that the functional groups in these materials promote the adsorption of plasminogen and its activation to a plasmin-like molecule. It appears from the substantial quantities of fibrinogen adsorbed to these materials after 3 h exposure to plasma that the Vroman effect (giving transient adsorption of fibrinogen) is not operative on these materials. It is hypothesized that specific interactions occur between fibrinogen and sulfonate groups.
INTRODUCTION
Insoluble polystyrenes bearing sulfonate and amino acid sulfamide groups of various kinds (see Fig. 1) have been shown to interact specifically with different plasma proteins or serine proteases.',' Most of these resins exhibit heparin-like properties in contact with plasma, i.e., they exhibit a specific in-
To whom correspondence should be addressed. Journal of Biomedical Materials Research, Vol. 25, 67-84 (1991) CCC 0021-9304/91/010067-18$04.00 0 1991 John Wiley & Sons, Inc.
68
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
NH
1
CH-R1
1
COOR2
Figure 1. Structure of the modified polystyrene resins.
teraction with ATIII and thrombin. Resins bearing L-arginine methyl ester groups do not exhibit heparin-like activity, but rather exert a direct action on thr~mbin.~ It has been shown that different parameters have an influence on the protein interactions of such biomaterials, e.g., the type of amino acid, the amino acid content, and the ratio of free sulfonate to amino acid substituted sulfonate. Much of the mechanistic work reported thus far on these resins, has been done in buffered solutions of single proteins. But studies of protein adsorption from plasma generally show results considerably different from solutions of single proteins or relatively simple mixtures consisting of a few protein^.^ Quantities of individual adsorbed proteins are generally less from plasma and binding affinities may also be modified. Also dynamic effects involving displacement of one protein by another have been observed.5s In adsorption from plasma, differences in protein composition between the adsorbed layer and plasma have generally been found. Moreover, the composition of the layer varies from surface to surface and is believed to determine the ultimate response of the blood to contact with the material. Therefore it was of interest to study the protein layer composition, and to identify the proteins adsorbed from plasma onto the family of polystyrene resins described above. In particular, it was of interest to see if differences in protein layer composition existed among the different resins which could be related to their blood interactions. For this purpose, proteins adsorbed after 3 h of contact with plasma, were eluted sequentially by 1 M Tris buffer and 4% SDS. SDS polyacrylamide gel electrophoresis (SDS-PAGE) of eluted proteins was performed and showed a multiplicity of components in the eluates from all resins. Positive identifications were made for one resin by immunodiffusion against specific antibodies. These results were correlated to the presence of different chemical groups and to the amino acid and sulfonate content of the resins.
PROTEINS IN HUMAN PLASMA
69
MATERIALS A N D METHODS
Modified polystyrene resins L-Arginyl methyl ester sulfamide polystyrene (PAOM), L-seryl methyl ester sulfamide polystyrene (PSerOM), and sulfonated polystyrene (PSSO3) were prepared according to a method previously described.' L-Glutamic acid sulfamide polystyrene (PSS02Glu) was kindly provided by Dr. C. Fougnot.' The compositions of the resins are given in Table I. The starting material consists of crosslinked polystyrene beads of diameter 37-74 pm, (Biobeads SX2, Bio-Rad, Richmond, CA, USA) and is used as a control (PS). The resins were equilibrated in isotonic Tris buffer pH 7.4.
Proteins Human fibrinogen was obtained from Kabi (Stockholm, Sweden). The lyophilized product was dissolved in distilled water and dialyzed overnight against an appropriate buffer, usually 0.05 M Tris pH 7.4, and stored frozen until used. This material is known to contain traces of immunoglobulin, a2 macroglobulin, and plasminogen (less than 0.2%). Other suspected contaminants are fibronectin and fibrinogen degradation products. SDS PAGE of this fibrinogen shows, after reduction with /3-mercaptoethanol,in addition to the expected bands for the a, /3, and y chains of fibrinogen (molecular weights (MWs) of 67,56, and 53 kD, respectively)additional bands at higher and lower MW consistent with the above mentioned contaminants (see Fig. 3). Fibrinogen degradation products (FDP) were obtained by treatment of Kabi fibrinogen with streptokinase. After various times, €-amino caproic acid was added to quench the reaction. Human plasminogen was prepared in our laboratory" and kindly provided by S. Khamlichi. Human albumin was from Behringwerke (Marburg, West Germany) and was 100% immunochemically pure. Porcine heparin was obtained from McMaster University Hospital TABLE I Chemical Composition of Polystyrene Resinsa Amino Acid Substituted Benzene Rings ~
Unsubstituted Benzene Rings
Sulfonated Benzene Rings
PSSOS PAOM5
10,o 14,5
90,O 80,5
-
-
-
5,o
(CH213 NH
CH3
PAOM71 PSS02Glu PSerOMlO
19,O 22,5 14,5
10,o
71,O 3,6 10,o
NH,-C=NH (CH&-COOH CHzOH
CH3 H CH3
73,9 75,5
R1
"Data are given as percent of benzene rings in specified form.
I
R2
70
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
Hamilton, Canada. Agkistrodon Rhodostromm (snake venom, Arvin) was kindly provided by Dr. M.W C. Hatton (McMaster University). Sepharose 6B was from Pharmacia, Ltd (Montreal, Canada). Eagle’s Minimum Essential Medium (MEM) was bought from Gibco (Burlington,Ontario, Canada). Antisera to human plasma proteins were from the following sources: anti haptoglobin, anti C3, anti Fc fragment of IgG, anti antithrombin 111, anti a2 macroglobulin, anti fibronectin, anti IgM and anti a1 antitrypsin were from Cappel Cochranville, PA, U.S.A. Anti IgG (H & L chains) and antilipoproteins were from Miles Scientific, Rexdale, Ontario, Canada.
Blood and plasma Human platelet poor plasma (PPP) was prepared from acid-citrate dextrose anticoagulated blood (6 vol blood: 1 vol ACD solution). The whole blood was taken from normal, healthy, single donors into citrate anticoagulant and centrifuged (3 min 12008, 3 min 1200g, 15 min 25008 at 20°C). The plasma was then stored at -70°C. Arvinized plasma was prepared as previously described’ from fresh human blood collected into ACD.
Heparinized plasma The heparin was contained in vials, each containing 100 anticoagulant IU/mL in isotonic saline solution. To allow for easier mixing into the blood, the heparin was diluted with isotonic Tris buffer to 15 mL. It was then mixed with 150 mL of freshly drawn blood (single donor), and processed in the usual way to obtain platelet poor plasma.
Electrophoresis and gels Acrylamide, bisacrylamide, SDS, silver nitrate, Coomassie blue R250 and other electrophoresis products were from Bio-Rad (Richmond, CA, USA). Molecular weight standards were the low and high MWs calibration kits from either Pharmacia (Uppsala, Sweden), Bio-Rad, or Sigma (St. Louis, MO, USA). Sepharose-heparin was prepared using the CNBr method of Miller Andersson et al.” Double diffusion immunoassay: Ouchterlony double diffusion immunoassays were performed on the samples eluted from the PAOM5 resin by 1M Tris buffer. Both antigens and antisera were placed into wells punched out of Agarose gel (l%), and were allowed to diffuse toward each other.” Following diffusion, the gels were stained with 1.5%Naphtol Blue Black B solution.
PROTEINS IN HUMAN PLASMA
71
Adsorption procedures
Four grams of resin particles suspended in isotonic Tris buffer pH 7.4, was packed into a glass column 17.5 cm x 1.3 cm. Ten milliliters of the plasma to be studied was then loaded onto the column. The system was incubated for various times from zero (direct elution) to 3 h at room temperature. The latter is a period sufficiently long that adsorption should have reached "equilibrium." The plasma was then removed by washing with isotonic Tris buffer until the adsorbance at 280 nm of the effluent returned to zero. The column was then eluted sequentially with 1M Tris pH 7.4, and with 4% SDS in isotonic Tris buffer. Eluted samples were concentrated by ultrafiltration (Amicon model 8010, 10 mL) and studied by SDS PAGE using 11% and 5 to 15% gradient gels.I3Samples were reduced by boiling in p-mercaptoethanol. Following electrophoresis, the gels were stained for protein with 0.05% Coomassie brilliant blue solution. Silver staining was used for selected cases.
RESULTS
Exposure of resins to PPP
The characterization of proteins adsorbed from plasma to polystyrene resins was performed after 3 h of contact at room temperature. A typical chromatogram from a column experiment is shown in Figure 2. Following
t1
iISOTONIC
I /I
TRlS
4%
SDS
100
FRACT ION NUMBER
b
Figure 2. Typical chromatogram for elution of ACD plasma from PAOM 5 resin.
72
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
the wash step with isotonic Tris buffer, sharp peaks were eluted both by 1M Tris pH 7.4, and 4% SDS in isotonic Tris for the resins PS, PSSO3, PAOM5, PSS02Glu, and PSerOM10. To check whether significant amounts of protein remained on the column after treatment with SDS, elution with 4M urea was performed. Only a very small peak was generally eluted and contained mostly fibrinogen, FDP, and albumin. Different elution patterns were observed with PAOM71. When ACD plasma was incubated with this resin, it clotted immediately. This behavior was attributed to the high concentration of positive charge present in PAOM71 compared to PAOM5. We hypothesized that when ACD plasma was contacted with PAOM71, the citrate ions formed a complex with the positive charge leaving the calcium free, and thus causing the plasma to coagulate. However, when the resin was pretreated with 2M trisodium citrate (referred to PAOM7lc) in order to saturate the guanidyl groups, no clotting occurred, and sharp peak was eluted as for the other resins. Such a pretreatment, however, presumably masks the effect of the guanidyl groups. Thus in order to study the influence of the free guanidyl groups on protein adsorption, heparinized plasma was contacted with PAOM71 and the results were compared with those obtained in the same conditions with ACD plasma and PAOM7lc. No significant differences were detected (data not shown). Figure 3 shows the SDS PAGE from a typical experiment using normal ACD plasma (3 h contact time) on PAOM5. The protein staining bands were tentatively identified by their position in the gel in relation to molecular weight standards. An experiment using a shorter contact time (no incubation) was undertaken to investigate possible kinetic effects and almost identical results were obtained (data not shown). As can be seen in Figure 3, the 1M Tris peak, lanes 2 and 3 (the latter analyses the end of the 1M Tris peak) shows bands at 94 kD (A, probably plasminogen), 80 kD (B, transferrin), and 69 kD (C, albumin).14In addition, a group of bands centred at about 40 kD (D) and the band at 14 kD (E) may be attributed to plasmin-induced FDP (see lane 6). The SDS peak (lane 4) also shows bands corresponding to plasminogen (A), transferrin (B) and albumin (C). The principal component, however, is fibrinogen (see lane 5) and some is degraded (lane E). The albumin and the aa band of fibrinogen run at the same place on the gel. Figures 4 and 5 show, respectively, the gels of the 1M Tris and SDS eluates from PS (lane 8), PSSO3 (lane 7), PSS02Glu (lane 6), PSerOMlO (lane 5), PAOM5 (lane 4)and PAOM71 (lane 3). Gels of FDP and fibrinogen are shown for comparison (lanes 1 and 2). No proteins were eluted in the 1M Tris peak from the PS column. For the other resins, the 1M Tris peak shows major bands at about 160 kD (A, PAOM71), 123 kD (B, PAOM71 and PSSO3), 118 kD (C, PAOM5 and PSSO3), 110 kD (probably plasminogen (Pl) PAOM71, PAOM5, PSerOM10, and PSSO3), 100 kD (D, PAOM71, PAOM5, PSerOM10, and PSSO3), 80 kD (transferrin (Tr) PAOM71, PAOM5, and PSS02Glu), 67 kD (albumin (Alb), PAOM71, PAOM5, PSerOM10, PSS02Glu, and PSSO3), 67-56 and 53 kD (a,p and y chains of fibrinogen (Fg), PAOM71, PAOM5, and PSS02Glu), 53-49 and 43 kD (FDP, PAOM71, PAOM5, PSerOM10, PSS02Glu, and PSSO3), 28 kD
PROTEINS IN HUMAN PLASMA
73
Figure 3. SDS PAGE of proteins eluted from PAOM after incubation with ACD plasma (3 h contact time). Samples were reduced in 2% p-mercaptoethanol and loaded onto an 11%gel. Lane 1: MW standards. Lanes 2 and 3: 1M Tris eluate. Lane 4: SDS eluate. Lane 5: Kabi fibrinogen (control). Lane 6: fibrinogen degradation products produced by incubation of Kabi fibrinogen with streptokinase for 3 h.
(E) and 20 kD (F). In the SDS peak, the major bands are at 154 kD (A, PAOM5), 108 kD (B, PAOM5), 100 kD (Pl, PAOM5), 81 kD (Tr, PAOM71, PAOM5, PSerOM10, PSS02Glu, and PSSO3), 67,56, and 53 kD (Alb, all resins and Fg, PAOM71, PAOM5, PSerOM10, PSS02Glu, and PSSO3) and groups of band at 30 kD (C) and 20 kD (D) (all resins).14There are also a few weaker bands. The behavior of PAOM7lc, in which guanidyl groups were neutralized by treatment with citrate is shown in Figure 6. In comparison with the same
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
74 1
2
3
4
5
6
7
8
Figure 4. SDS PAGE (5-1572 gradient gel, reducing conditions). Comparison of 1M Tris eluates after incubation of ACD plasma on PS, PSSO3, PSS02Glu, PSerOM10, PAOM5, and heparinized plasma on PAOM71. Lane 1: fibrinogen degradation products produced by incubation of Kabi fibrinogen with streptokinase for 3 h. Lane 2: Kabi fibrinogen (control). Lane 3: PAOM 71. Lane 4: PAOM 5. Lane 5: PSerOM10. Lane 6: PSS02Glu. Lane 7: PSSO3. Lane 8: PS. P1: plasminogen, Tr: transferrin, Alb: albumin, Fg: fibrinogen, FDP: fibrinogen degradation product.
resin in which the free guanidyl groups are available (Figs. 4 and 5, lane 3) there are significant differences. In the 1M Tris eluate, Figure 6 shows less albumin (Alb) and more transferrin (Tr) than Figure 4 (lane 3). In the SDS eluate, Figure 6 shows more fibrinogen (Fg) and less albumin (Alb) than Figure 5 (lane 3). Fibrinogen appears to be the major component of the adsorbed layer on all the resins studied, with the exception of PAOM7lc and PSerOM10. In the case of PAOM5 and PAOM7lc, it is eluted mainly by SDS (Figs. 3 and 6) except for the reconditioned resins, in which case it is eluted by both Tris and SDS (Figs. 4 and 5).
PROTEINS IN HUMAN PLASMA 1
2
75
3
4
5
6
7
8
Figure 5. SDS PAGE (5-15% gradient gel, reducing conditions). Comparison of 4% SDS eluates after incubation of ACD plasma on PS, PSSO3, PSS02Glu, PAOM5, PSerOM10, and heparinized plasma on PAOM71. Lane 1: fibrinogen degradation products produced by incubation of kabi fibrinogen with streptokinase for 3 h. Lane 2: kabi fibrinogen (control). Lane 3: PAOM71. Lane 4:PAOM5. Lane 5: PSerOM10. Lane 6. PSS02Glu. Lane 7: PSSO3. Lane 8: PS. P1: plasminogen, Tr: transferin, Alb: albumin, Fg: fibrinogen, FDP: fibrinogen degradation products.
Part of the eluted fibrinogen is partially degraded as indicated by good correspondence between the three bands 53 kD, 49 kD, and 43 kD in both the 1M Tris and SDS peaks, and corresponding bands in streptokinase induced fibrinogen degradation products (Figs. 4 and 5, lane 1).No FDP are observed on PAOM71 (Fig. 4 lane 3, and Fig. 6).
Exposure of PAOM5 to arvinized plasma Since fibrinogen appeared to dominate the protein layer adsorbed from normal plasma to these resins, it was of interest to investigate adsorption
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
76
1
2
3
4
5
Figure 6. SDS PAGE (ll%,reducing conditions) of proteins eluted from precitrated PAOM71 after contact with ACD plasma (3 h contact time). Lane 1: MW standards. Lane 2: 1M Tris eluate. Lane 3: 4% SDS eluate. Lane 4:Kabi fibrinogen (control). Lane 5: fibrinogen degradation products produced by incubation of Kabi fibrinogen with streptokinase for 3 h.
from plasma which was depleted of fibrinogen. This experiment was performed after 3 h of contact with arvinized plasma from which fibrinogen had been partially removed by treatment with a snake venom clotting enzyme. The concentration of the remaining fibrinogen in this plasma was determined and was found to be about 0.3%of the initial value.’ Figure 7 shows that the main protein components are albumin and high molecular weight proteins, at about 200 kD (A) in the 1M Tris peak and 120 kD (B) in the SDS peak. Many other lower MWs components are also present. Fibrinogen is also detected (lanes Fg) in spite of the low residual concentration. This indicates a high affinity of the protein for the polymer surface.
PROTEINS IN HUMAN PLASMA
77
Figure 7. SDS PAGE (ll%,reducing conditions) of proteins eluted from PAOM5 after contact with Arvinized plasma (3 h contact time). Lane 1: MW standards. Lane 2: 1M Tris eluates. Lane 3: 4% SDS eluates. Lane 4: Kabi fibrinogen. Arrows indicated the 200 kD (in the 1M Tris peak) and 120 kD (in the SDS peak) bands.
To identify the proteins eluted by 1M Tris, Ouchterlony immunodiffusions of column eluates were run against antibodies to various plasma proteins. The results of these analyses are shown in Figure 8. Positive identifications were made for C3 (well l), Fc fragment of IgG (well 3), IgG (well 4), fibronectin (well 7), and IgM (well 8). Such assays were not done for fibrinogen, FDP, albumin, and plasminogen, since it seemed clear from the gels that positive results would have been obtained for these proteins. Tests for other proteins were negative, which may reflect a concentration of the protein below the detection limit of the method. On the basis of the positive results obtained, we can attribute the bands at 200 kD to fibronectin. Human plasma fibronectin is composed of two polypeptides (aand p chains) each with MWs of about 200 kD. SDS PAGE of C3 has shown two polypeptide chains after
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
78
Figure 8. Ouchterlony analysis of 1M Tris fractions eluted from PAOM5 resin after contact with Arvinized plasma. Central wells contains the antisera: 1, anti C3; 2, anti haptoglobin; 3, anti Fc Fragment of IgG; 4,anti IgG; 5, anti d-antitrypsin; 6, anti antithrombin 111; 7, anti fibronectin; 8, anti IgM; 9, anti a2 macroglobulin; 10, anti @lipoprotein. Peripheral wells: 1M Tris eluate diluted to various extents.
denaturation and reduction. The polypeptide chains a and p have been assigned respective MWs of 115 11.5 and 45 k 7 kDs.”
*
Incubation of resins with fibrinogen solution in presence of added plasminogen To provide supporting evidence for the hypothesis that surface-bound fibrinogen in the presence of plasminogen undergoes lysis, experiments were performed with fibrinogen solution (3 mg/mL) to which plasminogen was added (3 kg/mL, i.e., 0.1% (w/w) relative to fibrinogen). Gels of the eluted proteins are shown in Figure 9, which includes controls generated by incubation of fibrinogen with 0.1% (w/w) plasminogen in isotonic Tris buffer. The 1M Tris eluate shows essentially unaltered fibrinogen and plasminogen. The groups of bands near 30 kD probably represents a small amount of FDP. In
PROTEINS IN HUMAN PLASMA
79
Figure 9. SDS PAGE (10-15% gradient gel, reducing conditions) of proteins eluted from resins after contact with fibrinogen solution containing 0.1% plasminogen. (A) Lane 1: low-molecular-weight standard, lanes 2, 3, 4, 5, and 6 : 4% SDS eluates from PSEOM10, PAOM71, PAOM5, PSSO3, and PSSOZGlu, respectively. Fg: fibrinogen, FDP: fibrinogen degradation products. (B) Lane 1: low-molecular-weight standard, lanes 2, 3, 4, 5, and 6 : 1M Tris eluates from PSEOMlO, PAOM71, PAOM5, PSSO3, and PSS02Glu, respectively.
the 4% SDS peak, FDP is present and in an extensively degraded form. In contrast to the other resins, PSerOMlO appears to induce considerably greater and more extensive degradation of fibrinogen. Only a small quantity of intact fibrinogen is present (Fig. 9, lane 5). Table I1 shows data on the relative quantities of fibrinogen plus plasminogen retained and eluted from the different resins. As can be seen, the arginine substituted resins adsorb significantly smaller amounts of the proteins than all the other resins (1M Tris eluate). On all the resins studied the adsorbed proteins are eluted mainly by SDS.
DISCUSSION
The exposure of blood to artificial surfaces leads to a rapid and competitive adsorption of plasma proteins within a matter of s e c ~ n d s .The ~ ” ~composition of the adsorbed layer differs considerably depending on the nature of the material. As can be seen from the gels in Figures 3, 4, and 5, the protein mixtures eluted from the polystyrene resins used in this study are complex and the gels contain many bands. The results of Ouchterlony analysis (Fig. 9) also show the extreme complexity of the process of adsorption from plasma. In general a much smaller fraction of the proteins adsorbed are eluted by 1M Tris (0.5-7%) than by 4% SDS solution (5-18%) (Table 111). This observa-
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
80
tion may be interpreted as showing that most of the proteins are bound by interactions other than electrostatic. The gel patterns of proteins eluted following 3 h of plasma contact show certain similarities on all the resins except the "starting" unmodified polystyrene (PS). Bands corresponding to albumin, plasminogen, fibronectin, IgG, and other unidentified bands (160, 150, and 43 kD) are generally observed. Fibrinogen appears to be a major component in every case. This is perhaps surprising since although fibrinogen is a major protein component of blood and is known to have a high surface affinit~,'~ it has been shown that its adsorption from blood or plasma is transient.>* This is the well known "Vroman effect" by which more abundant, low-surface-affinity proteins are adsorbed initially and later displaced by less abundant, high-surface-affinity proteins, e.g., the contact phase clotting factors." Since there are still substantial quantities of fibrinogen adsorbed to the present resins after 3 h of plasma contact, it appears that the Vroman effect is less pronounced on these materials. In agreement with this conclusion, we have shown recently that the Vroman effect is greatly reduced on polyurethane materials containing sulfonate groups in the hard segments." It appears that specific interactions may occur between sulfonate groups and fibrinogen. Compared to the other resins, the unmodified polystyrene (PS) apparently shows much less adsorption. Perhaps more precisely, less protein is eluted from this resin. This material is strongly hydrophobic with a water contact angle of about 80" and presumably can bind proteins only through hydrophobic interactions. Hydrophobic interactions contribute favorably both to AH and to AS: to the former via van der Waals interactions, and to the latter through dehydration of the interacting surfaces involving a disordering of water molecules. This entropy gain is considered to represent a major driving force for protein adsorption on hydrophobic surfaces. As can be seen from Figures 4 and 5, lane 8, no proteins were eluted by 1M Tris while only a few (some albumin, fibrinogen and probably IgG) were eluted in 4% SDS. These results agree with the idea that electrostatic interactions do not occur on this surface. In addition, they may show that SDS can disrupt van der Waals interactions between proteins and polystyrene. It must be emphasized that proteins are known to be adsorbed to polystyrene but may be particularly strongly bound and difficult to elute esTABLE I1 Relative Quantities (Percent) of Proteins Eluted from Polystyrene Resins by Isotonic Tris, 1M Tris, and 4%SDS Buffers, Following 3 h Contact with Fibrinogen Solution (3 mg/mL) to Which Plasminogen Was Added (3 pg/mL)
PSSO3 PSS02Glu PAOM5 PAOM71 PSerOMlO
Isotonic Tris Buffer
1M Tris
4% SDS
4 10 59 45 4
16.0 25.5 2.0 2.0 29.0
80.0 64.5 38.0 53.0 51.0
81
PROTEINS IN HUMAN PLASMA
TABLE 111 Relative Quantities (Percent) of Proteins Eluted from Reconditioned Polystyrene Resins by Isotonic Tris, 1M Tris, and 4% SDS Buffers, Following 3 h Contact with Plasma ~
PSSO3 PSS02Glu PAOM5 PAOM71 PSerOMlO
Isotonic Tris Buffer
1M Tris
4% SDS
80 89 88 89 75
7.0 5.5 0.5 6.0 7.0
13.0 5.5 13.5 5.0 18.0
pecially after long contact times. We showed previ~usly'~ that IgG, albumin, and fibrinogen adsorb in substantial quantities to polystyrene from both single protein solutions and from plasma. In addition, fibrinogen adsorption from plasma was not transient in contrast to many other surfaces,s8 again suggesting strong binding to polystyrene. More recently, Rapoza and Horbettz0 have shown that the SDS elutability of fibrinogen adsorbed on polystyrene is low compared to other surfaces. It therefore seems more likely that the relative lack of proteins in the eluates from PS is due to difficulty of desorption rather than to low adsorption. The substitution of sulfonate groups in the benzene rings of polystyrene causes a significant increase in hydrophilicity, leading to much greater adsorption of proteins from plasma, or easier elution of adsorbed proteins, notably fibrinogen. On conversion of some of the sulfonate groups to amino acid sulfamide groups, the protein adsorption behavior in plasma does not appear greatly altered. However, there are some important differences among the different amino acid or amino acid ester substituents. Thus, for the highly substituted arginine methyl ester resins, there appears to be less adsorption of proteins elutable in SDS (see Table 111). This result may be related to the relatively small concentration of sulfonate groups in PAOM71 resin. For this resin an experiment was also performed in which the guanidyl groups were "blocked" by treatment with trisodium citrate prior to ACD plasma exposure (see Fig. 6). In the 1M Tris peak, very few proteins are present, and there is much less albumin and virtually no fibrinogen. However a significant amount of plasminogen is observed. The SDS peak contains mostly transferrin, fibrinogen, and possibly some FDP. In general for the PAOM71 resin, whether the guanidyl groups are blocked or free, the generation of FDP is not extensive. It thus seems likely that sulfonate groups play a major role in fibrinogen degradation via plasminogen activation on these resins, as shown by Kichenin-Martin et a1." These authors studied adsorption of plasminogen on PSS02Glu and PSSO3. They have observed that purified plasminogen is adsorbed and becomes enzymatically active on the surface of these modified polystyrenes without any cleavage of the molecule. (Plasminogen is a single-chain protein of MW 94 kD and is normally converted to plasmin by enzymatic cleavage of a peptide bond within a double
82
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
disulphide loop.22This creates an enzymatically active molecule having a light chain of MW 25 kD and a heavy chain of MW 60 kD held together by disulfide bonds. The active site resides on the light chain.) A similar mechanism has been observed for the streptokinase/plasminogeninteraction and involves the formation of a plasminogen-streptokinase complex possessing enzymatic activity similar to that of pla~min.’~ In previous we showed that when glass beads were contacted with purified fibrinogen to which a trace of plasminogen was added, degradation of eluted fibrinogen was very extensive. These results were also interpreted in terms of a surface-mediated activation of plasminogen to plasmin followed by fibrinogenolysi~.~~ In connection with surface “fibrinolytic activity,” taken as the ability of a surface to adsorb and activate plasminogen, the serine methyl ester modified resin (PSerOM10) is of particular interest, and appears to have strong fibrinolytic properties as revealed by the experiments with fibrinogen-plasminogen mixtures (Fig. 9 and Table 11). The mechanism by which this activity is expressed is, for the moment, unknown and may not be the same as for the sulfonate-mediated effect (see above). It is perhaps relevant to point out that the methyl ester of serine is somewhat analogous to the seryl site of serine proteases which converts proenzymes to enzymes, for example, the coagulation factors. Thus, it is possible that plasminogen is converted to plasmin on this surface by the conventional cleavage mechanism.25Further work would be required to investigate this possibility. Most of the plasma data were obtained after 3 h incubation. However, it seemed possible that different adsorption patterns would be seen at different times. Indeed, adsorbed protein layers are known to exchange with proteins in solution and thus the composition of the layer may change over time. Therefore, an experiment was performed on one resin, PAOM5, in which the plasma was passed directly through the column without incubation (data not shown). The adsorption patterns were, in general, similar to those observed at longer contact time. However, it was noted that more intact fibrinogen was present after the short contact time. This finding is in agreement with the idea that, via the Vroman effect, less fibrinogen displacement should have occurred at shorter contact times. Surfaces which retain fibrinogen more strongly, like PSSO3, PSS02Glu and PSerOMlO (Table 111), may inhibit surface-induced coagulation by preventing access of coagulation factors to the surface. But they may, by the same token, be more platelet reactive. In this connection, however, ex vivo animal experiments performed on small-diameter tubings made of polystyrene-polyethylene copolymers grafted with sulfonate and aspartic acid sulfamide groups showed no significant platelet adhesion and aggregation on the surface of the tubings, during the 2 h of the experiment.26This was true whether the tubing was pretreated with either plasma or antithrombin 111. The resins used in this study were shown to interact specifically with thrombin and AT I11 in previous experiments,’ the specific affinity of the two proteins being dependent upon the properties of the resin. From the present studies, however, no clear differences in protein layer composition after
PROTEINS IN HUMAN PLASMA
83
plasma contact existed among the different resins. The slight variations in adsorption from plasma were in terms of relative concentrations rather than type of protein. Sulfonate groups are clearly a factor contributing to the adsorption of the fibrinogen. They may also be involved in adsorption of plasminogen and in the surface-mediated activation of plasminogen to plasmin. Financial support of this work by the Heart and Stroke Foundation of Ontario, the Medical Research Council of Canada, the Natural Sciences and Engineering Research Council of Canada, and the Centre National de la Recherche Scientifique (CNRS UD 0502) of France is gratefully acknowledged.
References 1.
2.
3. 4. 5. 6.
7. 8. 9.
10. 11. 12. 13. 14. 15. 16.
C. Boisson, J. Jozefonvicz, and J.L. Brash, “Adsorption of thrombin from buffer and modified plasma to polystyrene resins containing sulphonate and sulphamide arginyl methyl ester groups,” Biomaterials, 9, 47-52 (1988). C. Fougnot, M. Jozefowicz, and R. D. Rosenberg, ’Adsorption of purified thrombin or antithrombin I11 on two insoluble anticoagulant polystyrene derivatives: I1 Competition with the other plasma proteins,” Biomaterials, 5, 89-93 (1984). C. Boisson, D. Gulino, and J. Jozefonvicz, ’Affinity of antithrombin I11 for insoluble modified polystyrene,” Thromb. Res., 46, 101-108 (1987). J. L. Brash and T. A. Horbett, Proteins at Interfaces, ACS Symposium Series, vol. 343, Amer. Chem. SOC.,Washington D.C., 1987. L. Vroman, A. L. Adams, G.C. Fischer, and P.C. Munoz, ”Interaction of high molecular weight kininogen, factor XI1 and fibrinogen in plasma at interfaces,” Blood, 55, 156-159 (1980). L. Vroman and A. L. Adams, ”Identification of rapid changes at plasmasolid interfaces,” J. Biomed. Muter. Res., 3, 43-67 (1969). T.A. Horbett, “Mass action effects on competitive adsorption of fibrinogen from hemoglobin solutions and from plasma,” Thromb. Haemostas., 51, 174-181 (1984). J. L. Brash and P. ten Hove, “Transient adsorption of fibrinogen on foreign surfaces: similar behavior in plasma and whole blood,” J Biomed. Muter. Res., 23, 157-169 (1989). C. Boisson, D. Gulino, and J. Jozefonvicz, ”PolystyrPnes insolubles modifies par des derives de l’arginine: syntheses, interactions, avec les proteines plasmatiques,” Bull. la Rev. SOC.Chim. France, 4, 84-87 (1985). D.G. Deutsch and E.T. Mertz, ”Plasminogen, purification from human plasma by affinity chromatography,” Science, 170, 1095-1096 (1970). M. Miller-Anderson, H. Borg, and L.O. Anderson, “Purification of ATIII by affinity chromatography,” Thromb. Res., 5, 439-442 (1974). 0. Ouchterlony, Handbook of lmrnunodiffusion and Immuno-electrophoresis,“ Ann Arbor Science Publishers, Ann Arbor, MI, 1968. U. K. Laemmli, ”Cleavage of the structural proteins during assembly of the head of the bacteriophage T4,” Nature, 227, 680-685 (1970). J. L. Brash, S. Uniyal, B. M. C. Chan, and A. Yu, “Fibrinogen glass interactions: a synthesis of recent research,” ACS Symp. Ser., Mars (1983). B.F. Tack and J.W. Prahl, “Third component of human complement: purification from plasm, a and physicochemical characterization,” Biochemistry, 15,4513-4521 (1976). L. Vroman, A. L. Adams, M. Klings, G.C. Fischer, P.C. Munoz, and R. P. Solensky, ”Reactions of formed elements of blood with plasma proteins at interfaces,” Ann. NX Acad. Sci., 283, 65-76 (1977).
BOISSON-VIDAL, JOZEFONVICZ, AND BRASH
84
17. S. Uniyal and J.L. Brash, ”Patterns of adsorption of proteins from human plasma onto foreign surfaces,” Thromb. Huemostas., 47, 285-290 (1982). 18. J. L. Brash, C. F. Scott, P. ten Hove, P. Wojciechowski, and R.W. Colman, “Mechanism of the transient adsorption of fibrinogen from plasma to solid surfaces: role of the contact and fibrinolytic systems,” Blood, 71, 932-939 (1988). 19. J. P. Santerre, N. Van Der Kamp, and J. L. Brash, “Effect of sulfonation of segmented polyurethanes on the transient adsorption of fibinogen from plasma: possible correlation with anticoagulant behavior,” Trans. SOC. Biomat., 12, 113 (1989). 20. R.P. Rapoza and T.A. Horbett, “Changes in the SDS elutability of fibrinogen adsorbed from plasma to polymers,” J. Biomater. Sci. Polym. Edn, 1, 99 (1989). 21. V. Kichenin-Martin, L. Pejaudier, and M. Steinbuch, “Unusual, activation mechanism of plasminogen,” Thromb. Res., 52,469-478 (1988). 22. H. R. Lijnen and M. D. Collen, Sem. Thromb. Huemostas., 8, 2-10 (1982). 23. L. A. Schick and F. J. Castellino, “Direct evidence for the formation of an active site in the plasminogen moiety of the streptokinase-human plasminogen activator complex,” Biochem. Biophys. Res. Comm., 57,47-52 (1974). 24. J, L. Brash, B. M.C. Chan, P. Szota, and J. A. Thibodeau, “Degradation of adsorbed fibrinogen by surface generated plasmin,” J. Biomed. Muter. Res., 19, 1017-1029 (1985). 25. E Bachman, ”Fibrinolysis,” in XIth Congress Thrombosis Haemostasis Brussels, 227-265 (1987). 26. V. Migonney, “Fonctionralisation de la surface interne de matCriaux tubulaires. Etude de l’inhibition de la thrombine par l’antithrombine 111?I la surface de ces matCriaux,” ThPse dEtat, Paris XIII, 4 June 1986.
Received January 10, 1990 Accepted July 13, 1990
