Evaluation of leukocyte adhesion on polyurethanes: the effects of shear stress and blood proteins

Biomaterials 21 (2000) 2295}2303

Evaluation of leukocyte adhesion on polyurethanes: the e!ects of shear stress and blood proteins Weiyuan John Kao* School of Pharmacy and Department of Biomedical Engineering, 425 N. Charter Street, University of Wisconsin*Madison, Madison, WI 53706-1515, USA

Abstract Leukocytes are a central cell type in directing host in#ammatory and immune processes; thus, its response to biomaterials is extremely important in understanding material}host interaction. Blood contacting biomaterials may activate the complement cascade, thus promote leukocyte adhesion and activation to the biomaterial surface. We hypothesize that the extent of complementmediated leukocyte activation is modulated by the material chemical formulation and the presence of #uid shear stress. Medical-grade polyurethanes with or without 4,4-butyldiene bis(6-tert-butyl-m-cresol) antioxidant additives and a rotating disk system were utilized to study cell adhesion under a well-characterized shear stress "eld. Radioimmunoassay and ELISA were employed to assess the extent of complement activity. The results showed that adherent leukocyte densities decreased with increasing shear stress and that leukocyte adhesion was decreased signi"cantly further by the presence of the antioxidant in the polyurethanes. Cell adhesion under #ow conditions was abolished when complement C3 protein was depleted from the test medium. An increase in complement Factor H adsorption was observed at high shear region; however, no change in the complete complement activation was observed in the presence of shear stress as indicated by the protein S-terminal complement complex level. Based on these results, oligopeptides designed from C3a, C5a, and "bronectin were grafted onto a cell-nonadhesive polymer surface to probe the molecular mechanisms of leukocyte adhesion as mediated by protein}receptor complexation. The results showed that C3a-derived peptides mediated higher adherent macrophage density when compared to that mediated by C5a- and "bronectin-derived peptides.  2000 Elsevier Science Ltd. All rights reserved. Keywords: Complement; C3 complement protein; Rotating disc; Fibronectin; Polyethyleneglycol

1. Introduction The host in#ammatory reaction is a normal response to injury and the presence of foreign objects. The magnitude and duration of the in#ammatory process has a direct impact on the stability and compatibility of biomaterials, hence a!ecting the e$cacy of biomedical devices [1,2]. Leukocytes are a central cell type in directing host in#ammatory and immune processes; thus, its response to biomaterials is extremely important in understanding material}host interaction. Several characteristic leukocyte functions are identi"ed as critical events in the material}host interaction. First, leukocytes recognize adsorbed proteins on the biomaterial surface and may adhere onto the surface via several adhesion ligand}

* Tel.: #1-608-263-2998; fax: #1-608-262-3397. E-mail address: [email protected] (W.J. Kao).

receptor superfamilies [3]. This function could be modulated by the presence of active cytokines and growth factors [2]. However, the exact interrelationship among material physicochemical properties, adsorbed proteins, adherent cells, and in#ammatory cytokines and growth factors is not yet fully characterized. The interplay among these events and factors are complex and dynamic; hence, current research e!orts focus on elucidating the speci"c interaction between material physicochemical properties (i.e., surface and bulk chemistry, hydrophobicity, mechanical properties), adsorption isotherms of certain proteins (i.e., "bronectin, "brinogen, albumin), and observed cellular behaviors (i.e., adhesion, giant cell formation, cytokine and growth factor release) [3}7]. Second, the process of adherent monocyte-derived macrophage activation and fusion to form multinucleated foreign body giant cells is unique to the macrophage phenotype. Foreign body giant cells have been demonstrated on biomaterials and the rate of material degradation underneath the giant cells has been

0142-9612/00/$ - see front matter  2000 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 0 0 ) 0 0 1 5 6 - 3

2296

W.J. Kao / Biomaterials 21 (2000) 2295}2303

shown to be markedly increased [8]. However, the molecular mechanisms involved in foreign body giant cell formation remain unclear. Third, activated leukocytes may release cytokines, growth factors, and other bioactive agents to modulate the function of other cell types in the in#ammatory milieu [1,3,9]. For example, a systemic rise in the level of macrophage-derived interleukin-1 may cause neutrophilia, an increased marrow cellularity, an increased nitric oxide concentration, and an increased hepatic acute phase protein synthesis. A local elevation in interleukin-1 concentration may increase neutrophil in"ltration, induce angiogenesis, increase antibody production, and increase lymphokine synthesis (i.e., interleukin -2, -3, -4, -5, -6, -7, -10, and -12) [9]. Polyurethanes are a unique class of thermoplastic elastomers, possessing superior physical properties when compared to thermoplastics and processing advantages when compared to conventional elastomers. Polyurethanes are composed of short, rigid urethane or urea sequences (crystalline hard segments) connected to long, #exible polyether chains (amorphous soft segments). The hard-to-soft segment ratio can be varied during synthesis to tailor polyurethane properties to the intended end use [10,11]. Good mechanical performance and biocompatibility have led to extensive medical applications for polyurethanes. Polyurethanes are utilized extensively in cardiovascular biomedical devices such as arti"cial hearts, AV-assist devices, pacemaker lead insulators, and vascular grafts. Our previous studies showed that the presence of antioxidant in polyurethanes a!ected leukocyte morphology, adhesion, phenotype development, and material biostability in vivo [12]. Although the e!ects of material physiochemical property on adherent cellular behavior are evident, the complex protein-mediated mechanisms of this interaction remain unclear. Furthermore, blood-contacting biomaterials are often subjected to a wide range of #ow conditions. Wall shear stress is an important factor in #ow-modulated host responses to biomaterials such as protein adsorption, leukocyte adhesion, and thrombosis/coagulation [12}19]. Hence, the role of #ow shear stress must also be considered in the elucidation of the dynamic interrelationship between cardiovascular materials and leukocyte behavior. The development of quantitative methodologies is crucial in delineating cell}material interactions. In vitro protocols allow investigators to address a speci"c cellular event under a controlled environment. In this paper, several in vitro techniques are presented to demonstrate means of understanding a part of the complex protein-modulated mechanisms involved in the interaction between cardiovascular biomaterials, leukocytes, and hemodynamics. A clear understanding of this interaction will provide insights into other issues related to cardiovascular materials such as device infection and thrombosis [18}20].

2. Shear stress modulates leukocyte adhesion Several in vitro systems were developed to mimic various #ow conditions in the study of the e!ect of various rheological parameters on the interactions between physiological phenomena and arti"cial surfaces. Such methods include cone-and-plate rheometer, channel #ow chamber, recirculation system, and stagnation point #ow model [16,21,22]. We had adopted a rotating disk system [23] as a model to generate a wide range of dynamic shear conditions to study the role of wall shear stress on leukocyte behavior. The rotating disk system o!ers several unique advantages: (a) well-de"ned variable shear stress levels can be induced in a single experiment, (b) classical rheological model of #ow characterization is applicable, (c) di!erent types of material sample can be employed within the system, and (d) individual components can be sterilized prior to each usage. The rotating disc system (Fig. 1) employed herein consists of a rotating arm connected to a sample holder. A polytetra#uoroethylene #uid medium vessel is placed in a temperature-regulated water bath. The system has negligible lateral motion and employs a solid-state controlled servo system driven by an electric motor with a variable speed range with $1% precision. A steadystate rotational velocity is achieved within one revolution. The #uid wall shear stress (q , dyn/cm) at the  surface of the rotating disk is proportional to the radial distance of the disk and is de"ned [12,23}26] as: q "0.8gr(u/l), where g (cPoise) is the viscosity of the  medium, r (cm) is the radial distance from the center of the disk, u (rpm) is the radial velocity of the rotating disk, and l (Stokes) is the kinematic viscosity of the medium. The viscosity of 20% human serum in RPMI-1640 is 1.343$0.036 cPoise (mean$s.e.m., n"4) and the measured kinematic viscosity is 0.01071 Stokes at 373C as

Fig. 1. Rotating disc system.

W.J. Kao / Biomaterials 21 (2000) 2295}2303

determined by using a cone-and-plate viscometer. To directly correlate adherent cell density and wall shear stress at the sample surface, the rate at which the cells are &delivered' to the surface must be independent of the radial distance of the disk. The #ux (J, cells/cm/s) of cells onto the disk is obtained from the Navier}Stokes equations and the convective-di!usion equations for the rotating disk in a Newtonian #uid [12,23,25,26]. The #ux is de"ned as J"0.62Dl\uC , where  D (cells/cm/s) is the di!usivity coe$cient of each cell and C (cells/ml) is the initial cell concentration in  the medium suspension. D is de"ned [26] as (k¹)/(6ngb), where k is the Boltzman constant, ¹ is the temperature, and b is the radius of a nonadherent cell. Based on the equation, the #ux is uniform at the sample surface and is independent of the radial distance from the center of the sample disk. Other forces exerted on the adherent cells on the sample material are centrifugal, gravitational, and pressure gradient of the medium [23,26]. These forces are insigni"cant when compared with the #uid shear force. For example, the ratio of centrifugal force to #uid shear force is 0.00126, indicating that the #uid shear force exceeds the centrifugal force by three orders of magnitude [12,26]. To maintain the validity of the above derivations, several conditions must be ful"lled. Zero slip, steady-state stress "eld, and in"nite medium are maintained by the speci"c vessel geometry and the location of the disk [23,28]. The Reynolds number at the disk boundary [12,23,25,27] is calculated to be 2470, indicating that the #uid dynamics is within the range of transitional lamellar #ow regime. The critical boundary layer at the disk [23,25] is 614 lm, which is approximately two orders of magnitude smaller than the disk radius. This ensures that the disk surface can be considered in"nitely large so that edge e!ects are negligible and the #ow behavior can be described analytically. Finally, the critical boundary layer is approximately two orders of magnitude larger than the diameter of an adherent cell, which ensures that the #ow "eld and the wall shear stress at the disk surface are not disrupted by the presence of adherent cells. The rotating disk system was employed to quantify the e!ects of wall shear stress on leukocyte adhesion. A steady-state rotational velocity of 350 rpm, which generates a range of wall shear stress of 0}19.5 dyn/cm from the center to the edge of the disk, respectively, was utilized for all experiments. This shear stress range encompasses most normal physiologic conditions [13,14,21]. The test samples are medical-grade polyurethanes with or without 4,4-butyldiene bis(6-tert-butyl-m-cresol) antioxidant additives [12,29] cut into disks measured 17 mm in diameter and gas sterilized with ethylene oxide before use. Fresh human blood-derived polymorphonuclear leukocytes (PMN) and monocytes were isolated per previously described procedure [12] from healthy adult

2297

donors. All cells were suspended at concentrations of 3;10 monocytes/ml or 3;10 PMN/ml in RPMI-1640 medium supplemented with 20% autologous serum at 373C. Under the condition described herein, the values for J were calculated to be 5.44 PMN/cm/s and 0.54 monocytes/cm/s. Samples were incubated with the cell suspension under #ow condition for 1 h, rinsed with PBS, and stained with Wright's stain. Adherent cells were measured using a computerized image analysis system coupled to a light microscope. The number of adherent cells was determined along the disk radius. Four 0.10 mm "elds spaced 903 apart were counted and summed. The "elds counted along the disk radius corresponded to shear stress ranges of 0}0.8, 2.6}3.4, 6.6}7.4, 9.6}10.4, 13.6}14.4, and 16.6}17.4 dyn/cm, beginning from the center to the edge of the disk. The adhesion index was used to normalize and determine the percentage of transported leukocytes that adhered to the surface and is de"ned as ([measured adherent cell density]/[Jt]);100, where t is the experiment time of 3600 s. The results (Table 1) showed that adherent PMN and monocyte densities on polyurethanes decreased rapidly in the region of approximately 0}3.4 dyn/cm and then decreased steadily thereafter with increasing shear stress. This trend was observed for both polyurethanes with or without the 4,4-butyldiene bis(6-tert-butyl-m-cresol) antioxidant additives when normal, nondepleted serum was used in the cell suspension medium. Signi"cantly lower adherent PMN and monocyte densities were observed on polyurethanes containing the antioxidant additives under all shear levels when compared with polyurethanes without the antioxidant additives. From scanning electron microscopy analysis, adherent PMNs and monocytes on both types of polyurethanes exhibited spherical morphology with minimal cytoplasmic spreading and few pseudopodial extensions at the shear stress of 0}0.8 dyn/cm region (Fig. 2a and c). At a shear stress of 16.6}17.4 dyn/cm, adherent PMNs and monocytes on polyurethanes without antioxidant additives were #attened with cytoplasmic spreading, extensive pseudopodial extension, and cell surface ru%ing (Fig. 2b). However, leukocytes on polyurethanes containing the antioxidant additives showed only minimal cytoplasmic spreading at all shear stress levels (Fig. 2a and d). These results indicate that the activation of adherent leukocytes increased with increasing shear stress but this behavior was impeded by the presence of antioxidants in the test sample. Furthermore, monocytes showed higher adhesion index when compared to PMNs at all shear stress levels on both types of polyurethanes. These adhesion index results, which normalize the adherent cell densities to the number of cells transported to the sample surface, indicate that monocytes had a greater propensity for adhesion than PMNs on polyurethanes under all shear levels.

2298

W.J. Kao / Biomaterials 21 (2000) 2295}2303

Table 1 Adhesion index for adherent PMNs and monocytes on polyurethanes (PU) under shear conditions Wall shear stress (dyn/cm)

Nondepleted serum PU without antioxidant

C3-depleted serum PU containing antioxidant

PU without antioxidant

FN-depleted serum PU containing antioxidant

PU without antioxidant

PMNs 0}0.8 2.6}3.4 6.6}7.4 9.6}10.4 13.6}14.4 16.6}17.4

640$60 280$40 100$30 90$20 60$20 40$10

360$120 10$10 10$0 0$0 0$0 0$0

0$0 0$0 0$0 0$0 0$0 0$0

0$0 0$0 0$0 0$0 0$0 0$0

100$50 30$20 10$10 10$10 10$10 10$10

Monocytes 0}0.8 2.6}3.4 6.6}7.4 9.6}10.4 13.6}14.4 16.6}17.4

2420$460
770$150
460$100
260$150
200$100
150$50


460$150
100$50
50$0
50$0
0#0 0$0

50$50 0$0 0$0 0$0 0$0 0$0

0$0 0$0 0$0 0$0 0$0 0$0

770$410
150$100
100$100
100$50
50$0
0$0

PU containing antioxidant

20$10 10$10 0$0 0$0 0$0 0$0 310$200
50$50
0$0 0$0 0$0 0$0

Antioxidant: 4,4-butyldiene bis(6-tert-butyl-m-cresol). Adhesion index is de"ned as [(measured adherent cell density)/(Jt)];100, (unit"%), where J is the cell #ux (5.44 PMN/cm/s, 0.54 monocytes/cm/s) and t is the time of the experiment (3600 s). All values expressed in mean$s.e.m., n"5.
Di!erences at 90% con"dence level (p(0.10) between monocytes and PMNs at the respective shear stress level, as determined by the Student's t-test. All values at 0}0.8 dyn/cm region are signi"cantly higher than corresponding values at other shear stress levels at 95% con"dence level (p(0.05).

3. Shear stress modulated complement-mediated leukocyte adhesion The complement cascade consists of various blood proteins that, upon activation by the presence of foreign objects, mediates the normal immune and host in#ammatory response [12,30}33]. Hence, blood-contacting biomaterials have the potential to activate the complement cascade upon exposure to blood components such as whole blood, serum, or plasma. For example, membranes used in extracorporeal hemodialyzers can activate the complement system and induce acute leukopenia commonly observed in dialysis patients [33]. The activation of the complement cascade by biomaterials occurs primarily through the alternative pathway (Fig. 3) with complement protein C3 being a critical component. C3 has been shown to adsorb onto biomaterial surfaces in a fragmented form called C3b. On an &activating surface', adsorbed C3b further complexes preferentially with complement Factor B in a series of events resulting in the formation of surface-bound C3b, C5b, and soluble C3a, C5a as a part of the alternative pathway. On an &inactivating surface', adsorbed C3b complexes preferentially with complement Factor H and Factor I forming the surface-bound iC3b, thus inhibiting the conversion of C5 and terminating further complement activity. The products of complement activation modulate various leukocyte functions and host in#ammatory responses. Surface-bound C3b or iC3b are

ligands for leukocytes via CD35 or CD11/CD18 and CD11c/CD18 cell-membrane receptors, respectively. Soluble C3a and C5a anaphylatoxins stimulate leukocyte chemotaxis, adhesiveness, and secretion of reactive oxygen intermediates and cytokines during acute in#ammation [31,33]. Surface-bound C5b may further bind sequentially to C6, C7, C8, and C9 to form the lytic transmembrane terminal complement complex, which may become solubilized and inactivated by binding to the plasma protein vitronectin to form protein S-TCC [33]. To determine the relative importance of the complement cascade in mediating the initial adhesion of leukocytes on polyurethanes under shear condition, a critical complement component, namely C3, was depleted from healthy human serum using anti-C3 antibody a$nity column. As a comparison, "bronectin was depleted from the serum using gelatin sepharose chromatography. No other irrelevant serum protein was depleted as a negative control. All protein depletion was con"rmed using SDS-PAGE and Western blot with polyclonal antibody against C3 or "bronectin [12]. Cell adhesion study under shear condition was investigated using the rotating disk system under similar conditions as described above. Culture medium was supplement with 20% normal nondepleted, C3-depleted, or "bronectin-depleted serum. The results (Table 1) showed that when C3-depleted serum was utilized, leukocyte adhesion on both

W.J. Kao / Biomaterials 21 (2000) 2295}2303

2299

Fig. 2. Scanning electron microscopy micrographs of adherent PMNs on polyurethanes at 0}0.8 dyn/cm shear stress region (a), at 16.6}17.4 dyn/cm shear stress region (b), and on polyurethanes containing 4,4-butyldiene bis(6-tert-butyl-m-cresol) antioxidant additives at 0}0.8 dyn/cm shear stress region (c), and at 16.6}17.4 dyn/cm shear stress region (d). Adherent cells exhibited spherical morphology with minimal cytoplasmic spreading and few pseudopodial extensions at the shear stress of 0}0.8 dyn/cm region (Fig. 2a and c). At a shear stress of 16.6}17.4 dyn/cm, adherent cells on polyurethanes without antioxidant additives were #attened with cytoplasmic spreading, extensive pseudopodial extension, and cell surface ru%ing (Fig. 2b). However, leukocytes on polyurethanes containing the antioxidant additives showed only minimal cytoplasmic spreading at all shear stress levels (Fig. 2c and d).

polyurethanes under shear conditions was completely abolished at all shear levels. These results indicate the relative importance of the complement cascade, speci"cally the presence of C3 complement protein, in mediating leukocyte adhesion on polyurethanes under dynamic #ow conditions. When "bronectin-depleted serum was used, leukocyte adhesion on polyurethanes without the antioxidant additives was signi"cantly (p(0.10) lower at all shear stress levels than that observed when nondepleted serum was employed. On polyurethanes with the antioxidant additives, leukocyte adhesion was practically absent even at low shear levels when "bronectin was depleted from the test medium. The adhesion index results further indicate that monocytes had a greater propensity for adhesion than PMNs on polyurethanes under shear condition when "bronectin was depleted from the test medium. To further investigate the role of complement proteins and "bronectin in modulating leukocyte adhesion on polyurethanes under #ow conditions, the levels of com-

plement factors and "bronectin on polyurethanes were quanti"ed. Soluble S-TCC was quanti"ed using a commercially available ELISA kit as an indicator of complete complement activation. Surface-bound C3, Factor B, Factor H, and "bronectin were quanti"ed using a previously developed radioimmunoassay technique based on protein G binding to the primary antibody [12]. The adsorption of these critical complement proteins indicates the propensity of the test samples for complement activity. Hence, both soluble and surface-bound indicators are necessary to obtain a reliable assessment of the activation of the complement cascade by the test surfaces. The radioimmunoassay technique employed herein is capable of detecting multiple surface-bound proteins from a multicomponent system. Furthermore, protein G binds to primary antibody with high avidity, thus eliminating the need for radiolabeling the primary antibody. Brie#y, polyurethanes without the antioxidant additives were incubated with medium supplemented with 20% normal, nondepleted serum. The samples were incubated for 1, 5,

2300

W.J. Kao / Biomaterials 21 (2000) 2295}2303

Fig. 3. The alternative complement pathway showing control proteins and leukocyte-active fragments.

30, and 60 min under dynamic shear conditions using the rotating disk system as described above. Samples were also incubated with the test medium in 24-well tissue culture polystyrene culture plates as static controls. The di!usion}convection #ux of proteins from a homogeneous protein mixture was considered comparable under static and shear conditions [12,26]; thus, results for individual proteins were comparable between static and dynamic shear conditions. At each incubation time, sample materials were washed with PBS, and nonspeci"c sites were blocked with 15% milk containing 20 lg/ml protein G. The samples were washed and cut radially into six equal pieces (i.e., one piece for each adsorbed protein of interest). Each piece was incubated with one primary antibody diluted in 1% ovalbumin for 12 h. The dilution for the primary antibodies to several human proteins had been determined previously [12]. Rabbit anti-a-fetoprotein was employed as a negative control. The samples were then incubated with I-protein G diluted in 1% ovalbumin, washed, dried, cut concentrically into two pieces to measure protein adsorption in low (0}10 dyn/cm) and high (10}19.5 dyn/cm) shear stress regions. The samples were counted in a gamma counter

and the values were normalized to the sample surface area. The e!ect of shear stress on complement activity was determined by comparing the results of obtained under shear and static conditions (Table 2). Polyurethanes at all shear levels adsorbed signi"cantly more Factor B up to 30 min than it had absorbed under static condition at 60 min. In the 10}19.5 dyn/cm regions, polyurethanes adsorbed signi"cantly more Factor H at 30 and 60 min than that under static condition for 60 min. These results suggest that shear stress initially increased the propensity for complement activation (i.e., an increased Factor B adsorption) of polyurethanes. Conversely, at a longer time, shear stress increased the propensity for complement down-regulation (i.e., an increased Factor H adsorption) of polyurethanes. No di!erence was observed in S-TCC levels between static and shear conditions after 1 h of culture, indicating that the increase in Factor B and Factor H adsorption, as a result of the shear stress, did not a!ect the complete complement activation. Within the shear stress range tested herein, the adsorption of C3 and "bronectin was found independent of the shear stress level. Protein adsorption isotherm from 1 to 60 min under shear conditions showed a rapid initial adsorption to an equilibrium plateau that was achieved for C3, Factor B, and "bronectin within 1 min at all shear stress levels. However, in the 10}19.5 dyn/cm region, Factor H adsorption on polyurethanes peaked at 30 min and decreased thereafter.

Table 2 Protein activity of polyurethanes without antioxidant additives under static and shear conditions Protein

Shear stress (dyn/cm)

1 min

5 min

C3

Static 0}10 10}19.5

* * * 140$38 91$20 143$69 124$42 102$34 134$35 161$33 281$87 206$106

Factor B

Static 0}10 10}19.5

* * * 6$9 29$6
35$12
34$3
11$4 32$9
40$8
44$25
11$3

Factor H

Static 0}10 10}19.5

* * * 40$8 26$18 48$4 48$17 55$21 75$6


28$20 23$5 60$11


Fibronectin

Static 0}10 10}19.5

* * 14$11 3$3 19$10 14$6

* 17$8 13$13

16$15 4$4 3$4

S-TCC

Static 0}19.5

* *

* *

* *

30 min

60 min

510$230 630$430

All values are expressed in mean [counts-per-minute minus values of no antibody controls]/mm of sample surface$s.e.m., n"5 except S-TCC values are expressed in mean [ng/ml]$s.e.m. n"5.
Di!erences (p(0.05) between static controls and respective values.

W.J. Kao / Biomaterials 21 (2000) 2295}2303

4. Probing the molecular mechanisms of complement- and 5bronectin-modulated leukocyte adhesion Based on the investigations described above, it is clear that C3 complement protein and plasma "bronectin play an important role in mediating leukocyte adhesion on polyurethanes under #ow conditions. However, the molecular mechanisms involved in C3- and "bronectinmodulated leukocyte adhesion is not clear. Oligopeptides were designed based on the functional structure of the native C3 and "bronectin proteins to address the function}structural relationship between target proteins and cell membrane receptors. The peptides were employed to probe the molecular mechanisms of ligand}receptor recognition, post-ligation signal transduction, and cellular phenotypic development. The candidate peptides examined corresponded to residues 63}77 of complement component C3a (C3a}) [34], complement component C5a anaphylatoxin agonist derived from the C-terminal sequences of the native C5a protein [35], and residues 1615}1624 of "bronectin containing the RGD tripeptide cell-binding sequence (Fibronectin}) [36]. The amino acid sequences of the peptides are as follows: C3a }, WWGKKYRASKLGLAR}OH; C5a inhibitory sequence, FKYb-cyclohexylAb-cyclohexylLdAR}OH; and "bronectin}, RGDSPASSP}OH. A polymer network containing monomethoxy polyethyleneglycol monoacrylate, acrylic acid, and trimethylolpropane triacrylate was designed and synthesized as substrates for peptide grafting [37]. We demonstrated that the polymer networks without grafted peptides or with inactive peptides were free from macrophage, "broblast, and endothelial cell adhesion associated with the adsorption of serum or cellularly secreted proteins for up to 2 weeks of culture in vitro [37]. An established murine cell-line IC-21 [38] was employed as a macrophage model. Cells were incubated with the test materials placed in a 24-well TCPS culture plate with culture medium supplemented with 20% human serum for 2 h (i.e., designated as culture time"0). At this time, nonadherent cells were removed and the adherent cell densities were measured thereafter at various predetermined time points using a computerized video analysis system coupled to a light microscope. The results (Table 3) showed that C3a}-grafted surfaces supported higher adherent IC-21 macrophage densities than surfaces grafted with no-peptide controls and C5a-derived peptide and "bronectin} peptides at all culture times. C3a receptors have been found on granulocytes but are poorly characterized. The possible mechanisms through which ligated C3a receptors may result in leukocyte adhesion include the upregulation of CD11b/CD18 for iC3b, the upregulation of interleukin-1 production which could result in the expression of multiple adhesion receptors [39,40], and the direct binding of C3a receptors with the surface-bound C3a.

2301

Table 3 Adherent IC-21 macrophage density on polyethyleneglycol-based networks grafted with peptides Culture time (h)

No peptide C3a} control peptide

C5a Fibronectin} inhibitory peptide sequence

1 2 4 12 24 48 72 96

2$2 3$3 2$2 2$2 2$2 0$0 0$0 0$0

11$7 4$3 4$4 3$1 0$0 1$1 0$0 0$0

37$22 32$16 59$28 35$26 30$12 21$3 59$19 66$9

25$19 14$9 14$3 11$9 13$12 14$14 22$10 20$14

Cells were incubated with the sample surfaces for 2 h at which time (culture time"0) nonadherent cells were removed and the adherent cell densities were monitored thereafter. All values are expressed in cells/mm, mean$s.e.m., n"4}8. No di!erences in adherent cell density were observed between values of no peptide controls and C5a inhibitory sequence at all corresponding culture time. Adherent cell densities on surfaces grafted with C3a} were signi"cantly higher (p(0.10) when compared with respective values on no peptide grafted controls at all corresponding culture time.

Our current results indicate that the immobilized C3aU mediates leukocyte adhesion via the direct complexation with C3a receptors. The binding strength of this complexation was su$cient to maintain cell adhesion for a long duration of time. C3a had been shown to adsorb onto anionic biomaterial surfaces [41,42]; thus, our system is, in essence, a model to study the interactions between soluble and adsorbed C3a with C3a receptors. Adherent cell density on networks grafted with "bronectin-derived peptides was higher (p(0.10) than that on control surfaces without grafted peptides at all incubation times. Fibronectin} is derived from the module 10 of the central cell-binding domain of the native "bronectin protein. Grafted "bronectin} may mediate cell adhesion through the direct complexation with VLA5, CD51/CD29, and a b cell membrane 4  receptors [36,43] via the RGD-binding motif. From our previous study where "bronectin} and RGD tripeptides were utilized for grafting onto the polymer networks [37], the results indicate that the #anking sequence present in "bronectin} is necessary for mediating optimal cell adhesion. To determine the speci"city between grafted peptides and cell membrane receptors, cells were pre-incubated with free peptides (0.25 or 0.0025 mg/ml at 373C for 30 min) then cultured with polymer surfaces grafted with respective peptides. Adherent cell density was measured thereafter. Few adherent cells ((5 cells/mm) were observed up to 168 h of culture independent of the free peptide concentration used during the cell pretreatment.

2302

W.J. Kao / Biomaterials 21 (2000) 2295}2303

This trend was observed for C3a-, C5a-, and "bronectinderived peptides. At 168 h of culture time, nonadherent cells were removed, networks were rinsed with PBS, and incubated with fresh, untreated cells. Adherent cells densities increased signi"cantly and were restored within 2 h to values comparable to those observed in the previous cell adhesion studies (Table 3). The adherent cell densities increased steadily with increasing culture time. These results con"rmed the speci"city between grafted peptides and cell membrane receptors in mediating cell adhesion. These studies indicate the relative importance of the biofunctional region of complement C3a and "bronectin proteins in mediating cell adhesion via speci"c ligand}receptor complexation. The major function of leukocytes is to mediate host immune and in#ammatory response against foreign objects. Due to this reason, a clear understanding of the complex interaction between leukocytes and biomaterials is crucial in the improvement of materials employed in the construction of cardiovascular biomedical devices. Several methodologies were utilized to examine the interrelationship between material chemistry, leukocyte behavior, and hemodynamics at protein and cellular levels under well-de"ned in vitro conditions. Based on these "ndings, oligopeptides were designed, synthesized, and grafted onto a system of polymer networks to probe the molecular mechanisms of receptor}ligand recognition in mediating leukocyte adhesion. Acknowledgements This work was supported in part by the National Institutes of Health Grants HL-09601, HL-25239, HL47300, and HL-33849. The author thanks Profs. J.M. Anderson, A. Hiltner, and Dr. S. Sapatnekar of Case Western Reserve University, Cleveland, OH and Prof. J.A. Hubbell of Swiss Federal Institute of Technology (ZuK rich). References [1] Anderson JM. Mechanisms of in#ammation and infection with implanted devices. Cardiovasc Pathol 1993;2(3):33S}41S. [2] Anderson JM. In#ammatory response to implants. ASAIO 1988;11:101}7. [3] Ziats NP, Miller KM, Anderson JM. In vivo and in vitro interaction of cells with biomaterials. Biomaterials 1988;9:5}13. [4] Rudolph R, Cheresh D. Cell adhesion mechanisms and their potential impact on wound healing and tumor control. Clin Plast Surg 1990;17(3):457}62. [5] Kao WJ, Zhao QH, Hiltner A, Anderson JM. Theoretical analysis of in vivo macrophage adhesion and foreign body giant cell formation on poly-dimethylsiloxane, low density polyethylene, and polyetherurethanes. J Biomed Mater Res 1994;28:73}9. [6] Kaplan SS, Basford RE, Mora E, Jeong MH, Simmons RL. Biomaterial-induced alterations of neutrophil superoxide production. J Biomed Mater Res 1992;26:1039}51.

[7] Ziats NP, Pankowsky DA, Tierney BP, Ratno! QD, Anderson JM. Adsorption of Hageman factor (factor XII) and other human plasma proteins to biomedical polymers. J Lab Clin Med 1990;116(5):687}96. [8] Zhao Q, Topham N, Anderson JM, Hiltner A, Lodoen G, Payet CR. Foreign-body giant cells and polyurethane biostability: in vivo correlation of cell adhesion and surface cracking. J Biomed Mater Res 1991;25:177}83. [9] Wewers MD, In: Cytokines in health and disease. New York: Marcel Dekker, 1992. p. 235}56, 327}52. [10] Stokes K, McVenes R, Anderson JM. Polyurethane elastomer biostability. J Biomater Appl 1995;9(4):321}54. [11] Coury AJ, Slaikeu PC, Cahalan PT, Stokes KB, Hobot CM. Factors and interactions a!ecting the performance of polyurethane elastomers in medical devices. J Biomater Appl 1988;3(2):130}79. [12] Kao WJ, Sapatnekar S, Hiltner A, Anderson JM. Complementmediated leukocyte adhesion on poly(etherurethane ureas) under shear stress in vitro. J Biomed Mater Res 1996;32:99}109. [13] Giddens DP, Yoganathan AP, Schoen FJ. Prosthetic cardiac values. Cardiovasc Pathol 1993;2:167S}77S. [14] Slack SM, Turitto VT. Fluid dynamic and hemorheologic considerations. Cardiovasc Pathol 1993;2:11S}21S. [15] Willems GM, Giesen PLA, Hermens WT. Adsorption and conversion of prothrombin on a rotating disc. Blood 1993;82: 497}504. [16] Lawrence MB, McIntire LV, Eskin SG. E!ect of #ow on polymorphonuclear leukocyte/endothelial cell adhesion. Blood 1987;70:1284}90. [17] Hammer DA, Tempelman LA, Apte SM. Statistics of cell adhesion under hydrodynamic #ow: simulation and experiment. Blood Cells 1993;19:261}77. [18] Slack SM, Cui Y, Turitto VT. The e!ects of #ow on blood coagulation and thrombosis. Thromb Haemostas 1993;70: 129}34. [19] Ruggeri ZM. Mechanisms of shear-induced platelet adhesion and aggregation. Thromb Haemostas 1993;70:119}23. [20] Sapatnekar S, Kao WJ, Anderson JM. Leukocyte}biomaterial interactions in the presence of Staphylococcus epidermidis: #ow cytometric evaluation of leukocyte activation. J Biomed Mater Res 1997;35:409}20. [21] Guidelines for Blood}Material Interactions by the Devices and Technology Branch of the Division of Heart and Vascular Diseases. National Heart, Lung, and Blood Institute, NIH Publication 80-2185, 1979. p. 68}77. [22] Petschek H, Adamis D, Kantrowitz AR. Stagnation #ow thrombus formation. Trans Am Soc Artif Int Org 1968;14:256}60. [23] Wang IW, Anderson JM, Marchant RE. Platelet-mediated adhesion of staphylococcus epidermidis to hydrophobic NHLBI reference polyethylene. J Biomed Mater Res 1993;27:1119}28. [24] Mohandas N, Hochmuth RM, Spaeth EE. Adhesion of red cells to foreign surfaces in the presence of #ow. J Biomed Mater Res 1974;8:119}36. [25] Levich VG. Physiochemical hydrodynamics. Englewood Cli!s, NJ: Prentice-Hall, 1962. p. 57}79, 127}36. [26] Benton ER. On the #ow due to a rotating disk. J Fluid Mech 1966;24:781}800. [27] Wang I, Anderson JM, Marchant RE. Staphylococcus epidermidis adhesion to hydrophobic biomedical polymer is mediated by platelets. J Infect Dis 1993;167:329}36. [28] Daily JW, Nece RE. Chamber dimension e!ects on induced #ow and frictional resistance of enclosed rotating disks. J Basic Engng 1960;3:217}31. [29] Kao WJ, Hiltner A, Anderson JM, Lodoen GA. Theoretical analysis of in vivo macrophage adhesion and foreign body giant cell formation on strained poly(etherurethane urea) elastomers. J Biomed Mater Res 1994;28:819}29.

W.J. Kao / Biomaterials 21 (2000) 2295}2303 [30] Mollnes TE. Biocompatibility: complement as mediator of tissue damage and as indicator of incompatibility. Exp Clin Immunogenet 1997;14(1):24}9. [31] Mollnes TE. Complement and biocompatibility. Vox Sang 1998;74(Suppl 2):303}7. [32] Janatova J, Cheung AK, Parker CJ. Biomedical polymers di!er in their capacity to activate complement. Complement In#amm 1991;8(2):61}9. [33] Johnson RJ. Complement activation during extracorporeal therapy: biochemistry, cell biology and clinical relevance. Nephrol Dial Transplant 1994;9(suppl 2):36}45. [34] Ember JA, Johansen NL, Hugli TE. Designing synthetic superagonists of C3a anaphylatoxin. Biochemistry 1991;30:3603}12. [35] Sanderson SD, Kirnarsky L, Sherman SA, Vogen SM, Prakash O, Ember JA, Finch AM, Taylor SM. Decapeptide agonists of human C5a: the relationship between conformation and neutrophil response. J Med Chem 1995;138(18):3669}75. [36] Humphries MJ, Komoriya A, Akiyama SK, Olden K, Yamada KM. Identi"cation of two distinct regions of the type III

[37]

[38]

[39] [40] [41]

[42] [43]

2303

connecting segment of human plasma "bronectin that promote cell type-speci"c adhesion. J Biol Chem 1987;262(14): 6886}92. Kao WJ, Hubbell JA. Murine macrophage behavior on peptidegrafted polyethyleneglycol containing networks. Biotechnol Bioengng 1998;59(1):2}9. Crawford EK, Latham PS, Shah M, Hasday JD. Characterization of tumor binding by the IC-21 macrophage cell line. Cancer Res 1990;50:4578}83. Hugli TE. Structure and function of C3a anaphylatoxin. Curr Topics Microbiol 1989;153:181}208. Ross GD. Complement and complement receptors. Curr Opin Immunol 1989;2:50}62. Cheung AK. Complement activation as index of haemodialysis membrane biocompatibility: the choice of methods and assays. Neph Dial Transplant 1994;9:96}103. Hakim RM. Complement activation by biomaterials. Cardiovasc Pathol 1993;2:187S}97S. Hynes RO. Fibronectin. New York: Springer-Verlag, 1990.