Relation between cell density and the secretion of von Willebrand factor and prostacyclin by human umbilical vein endothelial cells.

In this study, the relation between the density of human umbilical vein endothelial cells (HUVECs) cultured on TCPS and (crosslinked) collagen, and the secretion of von Willebrand factor (vWF) and prostacyclin (PGI2) was determined. Collagen was crosslinked using N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide (EDC) in combination with N-hydroxy-succinimide (NHS), resulting in a matrix containing 14 free primary amino groups per 1000 amino acid residues after crosslinking (E/N14C). HUVECs were seeded on E/N14C, non-crosslinked collagen (N-Coll) and fibronectin-coated TCPS at densities ranging from 2500 to 50,000 cells/cm2. After 1 day of culture, both basal and A23187-stimulated secretion of vWF (expressed per 1,000,000 cells) was considerably increased at low cell densities (i.e. below 5000 cells/cm2) on all substrates. Secretion of PGI2 gradually increased with decreasing cell densities below 10,000 cells/cm2. After 10 days of proliferation, cell numbers on all substrates exceeded 50,000 cells/cm2, irrespective of the seeding density. Concomitantly, the initial higher secretion of PGI2 and vWF at the lowest seeding densities was decreased after longer times of culture, to values comparable to those obtained for higher seeding densities.

Binding and release of basic fibroblast growth factor from heparinized collagen matrices.

Endothelial cell seeding is a promising method to improve the performance of small-diameter vascular grafts. Growth of endothelial cells seeded on the luminal surface of synthetic vascular grafts, coated with a matrix suitable for cell seeding (e.g. collagen), can be accelerated by local, sustained release of basic fibroblast growth factor (bFGF). In this study two potential matrices for in vivo endothelial cell seeding were studied with respect to bFGF binding and release: collagen crosslinked using N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide (EDC) and N-hydroxysuccinimide (NHS), as well as heparinized EDC/NHS-crosslinked collagen. bFGF binding was determined after incubation of circular samples (10 mm diameter) with 0.25 ml bFGF solution for 90 min. Immobilization of increasing amounts of heparin, also using EDC and NHS, to crosslinked collagen containing 14 free primary amino groups per 1000 amino acid residues (E/N14C) resulted in binding of increasing amounts of bFGF. A plateau in bFGF binding was observed for heparinized E/N14C containing approximately 2.0-3.0 wt% of immobilized heparin which was obtained using a molar ratio of EDC to heparin-carboxylic acid groups of 0.4 during heparin immobilization (E/N14C-H(0.4)). At concentrations up to 840 ng bFGF/ml, 10% of the added bFGF bound to E/N14C, while binding of bFGF to E/N14C-H(0.4) amounted to 22%. Both E/N14C and E/N14C-H(0.4) pre-loaded with bFGF showed sustained bFGF release. A burst release of 30% in endothelial cell culture medium (CM) was observed for E/N14C during the first 6 h, compared to 2% release from E/N14C-H(0.4). After 28 days, the bFGF release from E/N14C and E/N14C-H(0.4) in CM amounted to 100 and 65%, respectively. Combined results of binding and release of bFGF indicate that compared to E/N14C, E/N14C-H(0.4) is the substrate of choice for bFGF pre-loading and subsequent endothelial cell seeding.

In vivo biocompatibility of carbodiimide-crosslinked collagen matrices: Effects of crosslink density, heparin immobilization, and bFGF loading.

Collagen matrices, crosslinked using N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide (E) and N-hydroxysuccinimide (N), were previously developed as a substrate for endothelial cell seeding of small-diameter vascular grafts. In the present study, the biocompatibility of various EN-crosslinked collagen matrices was evaluated following subcutaneous implantation in rats for periods up to 10 weeks. The effects of the crosslink density, referred to as the number of free primary amino groups per 1,000 amino acid residues (EN10, EN14, EN18, or EN22), the amount of heparin immobilized to EN14, and the effect of preloading heparinized EN14 with basic fibroblast growth factor (bFGF) on the induced tissue reaction were studied. EN-crosslinked collagen was biocompatible at both early and late time intervals, and matrices with high crosslink densities (i.e., EN14, EN10) especially demonstrated a significantly decreased antigenic response when compared to noncrosslinked collagen. Furthermore, increased crosslinking resulted in a decreased degradation rate. Immobilization of heparin onto EN14 resulted in a similar to EN14 (thus without heparin) or somewhat reduced tissue reaction, but fibrin formation and vascularization were increased with increasing quantities of immobilized heparin. Matrices preloaded with bFGF also demonstrated good biocompatibility, especially in combination with higher amounts of immobilized heparin. The latter matrices [EN14 with high heparin and bFGF, thus EN14-H (0.4)F and EN14-H(1.0)F] demonstrated significantly increased vascularization for periods up to 3 weeks. Neither heparin immobilization nor bFGF preloading induced an increased antigenic response. It is concluded that the results of this study justify further evaluation of bFGF preloaded, heparin immobilized EN14 collagen, as a matrix for endothelial cell seeding in experimental animals.

Immobilization of heparin to EDC/NHS-crosslinked collagen. Characterization and in vitro evaluation.

In the present study, heparin immobilization to a non-cytotoxic crosslinked collagen substrate for endothelial cell seeding was investigated. Crosslinking of collagen using N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide (EDC) and N-hydroxysuccinimide (NHS) resulted in a material containing 14 free primary amino groups per 1000 amino acid residues (E/N14C). At a fixed molar ratio NHS:EDC of 0.6, the amount of heparin covalently immobilized to E/N14C increased with increasing molar ratios of EDC to heparin carboxylic acid groups (Hep-COOH), to a maximum of approximately 5-5.5 wt% at a ratio of 2. Upon incubation in cell culture medium of endothelial cells, 4 to 7% of the immobilized heparin was released during 11 days. Immobilization of increasing amounts of heparin to E/N14C progressively reduced activation of contact activation proteases. Optimal anticoagulant activity, as measured by thrombin inhibition, was obtained after heparin immobilization using a ratio of EDC to Hep-COOH of 0.2-0.4 (14-20 mg heparin immobilized per gram of collagen). Platelets deposited to (heparinized) E/N14C showed only minor spreading and aggregation, although heparin immobilization slightly increased the number of adherent platelets. The results of this study suggest that heparin immobilization to EDC/NHS-crosslinked collagen may improve the in vivo blood compatibility of this material.

Endothelial cell seeding on crosslinked collagen: effects of crosslinking on endothelial cell proliferation and functional parameters.

Endothelial cell seeding, a promising method to improve the performance of small-diameter vascular grafts, requires a suitable substrate, such as crosslinked collagen. Commonly used crosslinking agents such as glutaraldehyde and formaldehyde cause, however, cytotoxic reactions and thereby hamper endothelialization of currently available collagen-coated vascular graft materials. The aim of this study was to investigate the effects of an alternative method for crosslinking of collagen, using N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide (EDC) in combination with N-hydroxysuccinimide (NHS), on various cellular functions of human umbilical vein endothelial cells (HUVECs) in vitro. Compared to non-crosslinked type I collagen, proliferation of seeded endothelial cells was significantly increased on EDC/NHS-crosslinked collagen. Furthermore, higher cell numbers were found with increasing crosslink densities. Neither the morphology of the cells nor the secretion of prostacyclin (PGI2), von Willebrand factor (vWF), tissue plasminogen activator (t-PA) and plasminogen activator inhibitor (PAI-1) was affected by the crosslink density of the collagen substrate. Therefore, EDC/NHS-crosslinked collagen is candidate substrate for in vivo application such as endothelial cell seeding of collagen-coated vascular grafts.

Endothelial cell seeding of (heparinized) collagen matrices: effects of bFGF pre-loading on proliferation (after low density seeding) and pro-coagulant factors.

Endothelial cell seeding to improve the performance of small-diameter vascular grafts requires a suitable substrate, such as crosslinked collagen. In addition to providing a suitable substrate for adhesion and growth of endothelial cells, proliferation of seeded endothelial cells can be enhanced by local, sustained release of basic fibroblast growth factor (bFGF, a heparin-binding growth factor for endothelial cells). We have previously shown that collagen crosslinked using N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide (EDC) and N-hydroxysuccinimide (NHS) supports adhesion and proliferation of human umbilical vein endothelial cells (HUVECs). In the present study, HUVECs were seeded on (heparinized) EDC/NHS-crosslinked collagen, pre-loaded with bFGF. Proliferation of HUVECs on (heparinized) crosslinked collagen increased with increasing amounts of pre-loaded bFGF. The minimal cell-seeding density required for proliferation proved to be very low after pre-loading the substrates with bFGF, and was 4-fold lower for heparinized crosslinked collagen compared to crosslinked collagen (250 versus 1000 cells/cm(2)). Pro-coagulant properties (von Willebrand factor secretion and tissue factor expression) of HUVECs seeded on (heparinized) crosslinked collagen, with or without pre-loading of bFGF, were comparable to those of HUVECs on TCPS. It is concluded that heparinized, EDC/NHS-crosslinked collagen pre-loaded with bFGF is a candidate matrix for in vivo endothelial cell seeding of synthetic vascular graft materials.

Improved endothelialization of vascular grafts by local release of growth factor from heparinized collagen matrices.

Endothelial cell seeding, a promising method to improve the performance of small-diameter vascular grafts, requires a suitable substrate, e.g. crosslinked collagen. In addition, the growth of seeded endothelial cells can be improved by local release of a heparin-binding protein, basic fibroblast growth factor (bFGF). In this study, the influence of immobilization of heparin to collagen, crosslinked using N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide (EDC) in combination with N-hydroxysuccinimide (NHS), on the binding and release of bFGF was determined. Heparin was immobilized also using EDC and NHS. Furthermore, the effects of the release of bFGF from (heparinized) EDC/NHS-crosslinked collagen on the proliferation of seeded endothelial cells was studied in vitro. Immobilization of increasing amounts of heparin to EDC/NHS-crosslinked collagen (containing 14 free epsilon-amino groups per 1000 amino acid residues, E/N14C) resulted in binding of increasing amounts of bFGF to the material. Maximal bFGF binding was observed for E/N14C containing 20-30 mg heparin immobilized per gram of collagen which was obtained using a molar ratio of EDC to heparin-carboxylic acid groups of 0.4 for heparin immobilization (E/N14C-H(0.4)). Up to concentrations of 320 ng bFGF/ml, 10% of the added bFGF bound to E/N14C, while binding of bFGF to E/N14C-H(0.4) was 22%. The initial release rate of bFGF bound to E/N14C was much higher compared to bFGF bound to E/N14C-H(0.4): respectively, 30 vs. 2% in the first 6 h. After 10 days, the bFGF release from E/N14C and E/N14C-H(0.4) amounted to 83 vs. 42%, respectively. Binding of increasing amounts of bFGF resulted in increased growth of human umbilical vein endothelial cells (HUVECs) seeded on both E/N14C and E/N14C-H(0.4). Nevertheless, after 6 and 10 days of proliferation cell numbers on E/N14C-H(0.4) where higher than cell numbers on E/N14C, irrespective of the bFGF concentration used for loading of the matrix. It is concluded that heparinized, EDC/NHS-crosslinked collagen is a good synthetic vascular graft coating for in vivo endothelial cell seeding.

Endothelialization of crosslinked albumin-heparin gels.

Crosslinked gels of albumin as well as heparinized albumin gels, potential sealants of prosthetic vascular grafts, were studied with regard to in vitro stability, binding of basic fibroblast growth factor (bFGF) and cellular interactions. A small percentage of the heparin present in these gels, was released during storage in SDS solution. During storage in cell culture medium at 37 degrees C, heparin release was 21-25 percent. Release of albumin did not occur. Human umbilical vein endothelial cells (HUVECs) rapidly adhered and subsequently spread on (heparinized) albumin gels, but proliferation was only observed if heparin was present in the gel. Binding of 125I-bFGF to heparinized albumin gel was 35 percent higher than to non-heparinized albumin gel. Growth of HUVECs occurred only on heparinized albumin gel loaded with bFGF and not on bFGF-loaded albumin gel. The number of platelets deposited under stationary conditions onto heparinized albumin gel was about twice the number found on nonheparinized albumin gel. Seeding of HUVECs on heparinized albumin gel, significantly reduced the number of platelets adhering to this surface. Moreover, no spreading of platelets was observed on substrates seeded with HUVECs. It can be concluded that crosslinked gels of albumin to which heparin is immobilized, are candidate sealants for prosthetic vascular grafts and suitable substrates for endothelial cell seeding.

Blood compatibility of surfaces with immobilized albumin-heparin conjugate and effect of endothelial cell seeding on platelet adhesion.

Endothelial cell (EC) seeding significantly improves the blood compatibility of artificial surfaces. Although a coating consisting of albumin and heparin (alb-hep) is a suitable substrate for seeded ECs, binding of ECs to the substrate further improves when small amounts of fibronectin are present in the alb-hep coating. Alb-hep conjugate was immobilized on carbon dioxide gas plasma-treated polystyrene (PS-CO(2)), thereby significantly increasing the recalcification time of blood plasma exposed to this surface. Furthermore, surface-immobilized alb-hep conjugate inhibited exogenous thrombin. Heparin activity was reduced by adding fibronectin on top of a monolayer of alb-hep conjugate, but not by simultaneous coating of fibronectin and alb-hep conjugate. Coating of PS-CO(2) with alb-hep conjugate significantly decreased contact activation (FXII activation). The number of platelets deposited from blood plasma on PS-CO(2) coated with alb-hep conjugate was twice as high as on PS-CO(2) coated with albumin. Addition of fibronectin to alb-hep conjugate-coated PS-CO(2) had no significant effect on the number of adhered platelets. Seeding of the substrates with ECs significantly reduced the number of adhered platelets under stationary conditions. Platelets deposited onto endothelialized surfaces were primarily found on endothelial cell edges, and sparingly on areas between ECs. In conclusion, alb-hep conjugate-coated surfaces display anticoagulant activity. ECs adhering to and proliferating on this coating significantly decrease the number of platelets which adhere to the surface. Therefore, alb-hep conjugate-coated surfaces form a suitable substrate for seeding of ECs in low density. Although application of fibronectin on top of the coating decreases the anticoagulant activity to some extent, it might be useful in view of the improved adherence of ECs to the coating.

Proliferation of endothelial cells on surface-immobilized albumin-heparin conjugate loaded with basic fibroblast growth factor.

Seeding of endothelial cells (ECs) on the luminal surface of small-diameter vascular grafts is a promising method to avoid occlusion of these prostheses. Immobilization of basic fibroblast growth factor (bFGF) to substrates used to coat or fill porous prostheses may enhance the formation of a confluent monolayer of ECs. Human umbilical vein endothelial cells (HUVECs) were grown on bFGF-loaded albumin-heparin conjugate bound to CO2 gas-plasma-treated polystyrene. In the order of 2-3 ng/cm2 bFGF had to be immobilized to form a confluent monolayer of HUVECs. The most prominent effect of surface-immobilized bFGF was stimulation of the proliferation shortly after seeding, resulting within 3 days in confluent cell monolayers with high density. In contrast, in cultures with 0.3 ng/mL bFGF in the medium instead of bFGF bound to the surface, it took almost a week before the cell layers reached confluency. Binding of bFGF to heparin and the biological activity of bFGF towards ECs were not influenced by the (radio-)labeling of bFGF with iodine. However, only a minor part of the bFGF used in this study displayed heparin affinity. Furthermore, degradation and multimerization of labeled bFGF in time occurred when the growth factor was stored at 20 degrees -37 degrees C. This limits the use of labeled bFGF to short-term (hours) experiments. In conclusion, bFGF loading of vascular graft surfaces through complexation of bFGF with a heparin-containing matrix probably will lead to more rapid formation of a confluent monolayer of ECs on graft surfaces upon seeding of the cells.

Adherence and proliferation of endothelial cells on surface-immobilized albumin-heparin conjugate.

Small-diameter vascular grafts rapidly fail after implantation, due to occlusion caused by thrombosis. This problem cannot be overcome using medication. A promising improvement of graft patency is the seeding of endothelial cells (EC) on the luminal surface of the vascular graft. Conjugates of albumin and heparin, which were developed to obtain nonthrombogenic coatings, could form an ideal coating for vascular grafts. Besides presenting anticoagulant function, heparin will bind proteins with cell adhesive properties, thus facilitating adherence of EC to the graft surface. EC were able to grow to confluency on CO(2) gas plasma-treated polystyrene (PS-CO(2)) coated with albumin-heparin conjugate. CO(2) gas plasma treatment resulted in the introduction of functional groups at the surface (e.g., hydroxyl, aldehyde, carboxylic acid, and epoxide groups). Addition of albumin-heparin conjugate to the functionalized surface in an aqueous solution with pH 8.2 yielded a stable monolayer of covalently bound conjugate. The number of cells adhering and proliferating on this surface was comparable to the number of cells on fibronectin-coated PS-CO(2). However, the structure and size of EC proliferating on surface-immobilized albumin-heparin was more irregular. Long-term adherence might be improved by adding fibronectin to the albumin-heparin surface, either as a mixture with albumin-heparin or in a separate incubation step.

Small-diameter vascular graft prostheses: current status.

In contrast to large-diameter vascular grafts (i.e. larger than 5 mm) which remain excellent for more than 10 years after implantation, small-diameter vascular grafts of both Dacron and Teflon occlude rapidly upon implantation. In this overview article, the strategies used to improve the patency of these small-diameter grafts, the current status in clinical trials, and further perspectives in the field of artificial vascular graft development are reviewed. It is concluded that, in view of recent developments in tissue engineering approaches, the future of small-diameter vascular prostheses looks promising.

Proteins involved in the Vroman effect during exposure of human blood plasma to glass and polyethylene.

The amounts of fibrinogen adsorbed to glass from various human blood plasmas have been measured as a function of time. The plasmas were 11 single donor plasmas, pooled plasma, a single donor high molecular weight kininogen (HMWK)-deficient plasma and HMWK-deficient plasma, which had been reconstituted with HMWK. For adsorption times between 1 min and 1 h more fibrinogen adsorbed from HMWK-deficient plasma compared with the amounts of fibrinogen which adsorbed from the other plasmas. This result supports the conclusion of several authors that HMWK is involved in the displacement of fibrinogen, initially adsorbed from normal human plasma to glass. Glass surfaces, pre-exposed to solutions of plasma and subsequently exposed to 1:1 diluted plasma, gives rise to a relatively high adsorption of HMWK which is independent of the plasma concentration of the precoating solution. The results indicate that HMWK from 1:1 diluted plasma is involved in the displacement of proteins from glass surfaces which had been pre-exposed to solutions with a low plasma concentration. Experiments with polyethylene as a substrate reveal that high density lipoprotein (HDL) from 1:1 diluted plasma is involved in the displacement of proteins from polyethylene surfaces which had been pre-exposed to solutions with a low plasma concentration. Moreover, evidence is presented that substantial amounts of albumin and fibrinogen, adsorbed from 1:1000 diluted plasma to glass and polyethylene, are displaced from the surfaces of these materials by proteins from 1:1 diluted plasma different from HMWK and HDL.

A mathematical model for the leukocyte filtration process.

Leukocyte filters are applied clinically to remove leukocytes from blood. In order to optimize leukocyte filters, a mathematical model to describe the leukocyte filtration process was developed by modification of a general theoretical model for depth filtration. The model presented here can be used to predict the time-dependent leukocyte filtration as a function of cell-cell interaction in the filter, filter efficiency, filter capacity, filter dimensions, and leukocyte concentration in the suspension applied to the filter. The results of different leukocyte filtration experiments previously reported in the literature could be well described by the present model. (c) 1995 John Wiley & Sons, Inc.

The binding of human blood platelets to fibrinogen-, fibronectin-, and Arg-Gly-Asp-derivatized Sephadex G-10.

The adhesive proteins fibrinogen (FG) and fibronectin (FN) were immobilized to glycine-Sephadex G-10. The derivatized Sephadex G-10 gels were used to bind human blood platelets. For comparison, Gly-Arg-Gly-Asp-Ser-Pro(GRGDSP)-derivatized Gly-Sephadex G-10 was used. FG-, FN-, and GRGDSP-Gly-Sephadex G-10 each bound a substantial number of activated blood platelets (> or = 5 x 10(8) ml-1 gel) while non-activated platelets were not bound. Binding of ADP-treated blood platelets to the affinity adsorbents was dependent on the ADP-concentration which was used, reaching a near-maximal value at about 10 microM ADP. Platelet binding to the three types of affinity gels could be completely inhibited by dissolved GRGDSP as well as monoclonal anti-platelet glycoprotein IIb/IIIa (GPIIb/IIIa) antibody CLB-C17, which demonstrates that platelet binding specifically involves the fibrinogen binding site on GPIIb/IIIa. Platelet binding to all three affinity gels required free Ca2+ and Mg2+ ions: platelets binding in the absence of these divalent cations was considerably lower than platelet binding in buffer containing 2 mM Ca2+ and 1 mM Mg2+. Moreover, activated ethylenediamine-tetraacetate (EDTA)-treated platelets did not bind at all to the affinity gels. The finding that non-activated platelets did not bind to the affinity gels is thought to be related to both the high hydrophilicity of the Sephadex basic material and to the native state of the gel-bound fibrinogen and fibronectin.

Effects of Tween 20 on the desorption of proteins from polymer surfaces.

The effects of Tween 20 on the desorption of proteins from polyethylene and polyurethane were studied, using single protein solutions of the human proteins fibrinogen (Fb), immunoglobulin G (IgG), serum albumin (HSA), high density lipoproteins (HDL), and plasma. The surfactant may partly or even completely desorb the proteins, depending on the type of polymer and protein. About 40% of adsorbed HSA and 80% of adsorbed HDL from the corresponding single protein solutions were desorbed by Tween 20 from polyethylene, whereas Tween 20 had a small effect on the desorption of adsorbed Fb and IgG under the same conditions. However, the desorption of Fb and IgG by Tween 20 was much higher in the case of a diluted plasma solution compared to a pure protein solution. These findings may be explained by the differences of the interaction strengths between polymers and the adsorbed proteins. The displacement of HSA from polyethylene by Tween 20 occurred in the first few minutes and did not increase in time. It was also observed that preadsorbed Tween 20 was able to prevent in a large extent the adsorption of HSA onto polyethylene. Thus, the effect of Tween 20 on the desorption of protein is due to either the displacement of protein or prevention of protein adsorption onto the surfaces.

Preparation of cell-affinity adsorbents by immobilization of peptides to Sephadex G-10.

In this study, affinity adsorbents for the binding of activated blood platelets and endothelial cells were prepared from Sephadex G-10 by immobilization of peptides, derived from the cell-adhesive amino acid sequence RGD (arginine-glycine-aspartic acid). Derivatization of Sephadex G-10 was accomplished by sequential coupling of specific dipeptides (RG and DV) (V = valine) and by coupling of the RGD-containing hexapeptide GRGDSP (S = serine, P = proline). Two types of gel were prepared by sequential coupling, designated as G10 (acetone) and G10 (dimethylformamide) (G10(DMF)), containing peptides which had been coupled to the outer side of the beads and throughout the porous beads, respectively. The binding capacity of the prepared Sephadex-derivatives amounted up to 2 x 10(9) human blood platelets per millilitre GRGDV-Sephadex at immobilized peptide concentrations, that were in the nanomole range (per millilitre packed gel) and which differed a factor 10 between G10 (acetone) and G10 (DMF). In a second series of experiments, different amounts of the hexapeptide GRGDSP were coupled to carboxylated Sephadex G-10 and carboxylated Sepharose CL 6B. The binding of human umbilical vein endothelial cells to the resulting materials was studied. Up to 10(6) endothelial cells attached per ml GRGDSP-derivatized hydrogel at peptide concentrations of 15 nmol GRGDSP/ml Sephadex and at +/- 300 nmol GRGDSP/ml Sepharose. Substitution of the arginine residue of the RGD-sequence by glutamine abolished the cell-binding activity of the immobilized peptide towards activated blood platelets but not towards endothelial cells. From the results of this study it can be concluded that small peptides can be coupled to the outer side of the porous Sephadex beads, resulting in high effective ligand densities for cell-affinity applications. In this respect, Sephadex G-10, derivatized according to 'the acetone method' is a good alternative for polystyrene and other solid phase materials.

Poly(ethyleneimine) modified filters for the removal of leukocytes from blood.

Polyurethane membrane filters and filters coated with poly(ethyleneimine) were used to investigate the influence of leukocyte adhesion during filtration. Treatment of the filters with an aqueous solution of 1% (w/v) poly(ethyleneimine) (PEI) led to the introduction of amine groups at the filter surfaces, as was confirmed by X-ray photoelectron spectroscopy. The modification procedure did not significantly change the porous structure in the filters, as was demonstrated by SEM and porometry. Using 14C-labeled poly(ethyleneimine) it was shown that nearly a complete coverage (approximately 0.1 mg/m2) was achieved that did not desorb from the filter surface during contact with blood plasma. When the filtration was carried out with purified leukocytes in the absence of red cells, platelets, and blood plasma, the number of cells removed by modified filters (> 95%) was significantly higher as compared to the removal with unmodified filters (approximately 80%). However, no significant differences between the filters were found when the filtration was performed with whole blood. This finding was unexpected, because it was shown before that immobilization of poly(ethyleneimine) on solid polyurethane film, surfaces promoted the adhesion of leukocytes from whole blood. Apparently, the adhesive properties of the PEI diminish during filtration. Filter coating of commercial leukocyte filters composed of polyester fibers also had no effect on the removal of leukocytes from whole blood. It was postulated that morphological factors, such as filter shape, roughness, tortuosity, and porosity rather than the physicochemical properties of the filter surface influence cell adhesion to the filter surface, and through that the filtration process.

In vitro leucocyte adhesion to modified polyurethane surfaces. I. Effect of ionizable functional groups.

To study the effect of ionizable functional groups on the adhesion of leucocytes to surfaces, both poly(ethyleneimine) and poly(acrylic acid) were immobilized on polyurethane films, resulting in the introduction of amine and carboxylic acid groups, respectively. This was confirmed by contact angle measurements and XPS analysis. In vitro adhesion of granulocytes and lymphocytes on untreated and modified surfaces was compared. The number of adherent cells on modified surfaces as a function of time was significantly higher than on untreated surfaces. This effect was most pronounced for the adhesion of lymphocytes to surfaces modified with amine groups. In this case, the number of adherent cells after 1 h of exposure was three times higher than on untreated surfaces. A moderate enhancement of leucocyte adhesion was observed in the case of surfaces modified with carboxylic acid groups. There is evidence that these groups were not ionized under the experimental conditions used. The modification procedures described may be used to improve polyurethane filters for the removal of leucocytes from blood.