Biomaterial surface chemistry dictates adherent monocyte/macrophage cytokine expression in vitro.

An in vitro human monocyte culture system was used to determine whether adherent monocyte/macrophage cytokine production was influenced by material surface chemistry. A polyethylene terephthalate (PET) base surface was modified by photograft copolymerization to yield hydrophobic, hydrophilic, anionic and cationic surfaces. Freshly isolated human monocytes were cultured onto the surfaces for periods up to 10 days in the presence or absence of interleukin-4 (IL-4). Semi-quantitative RT-PCR analysis on days 3, 7 and 10 of cell culture revealed that interleukin-10 (IL-10) expression significantly increased in cells adherent to the hydrophilic and anionic surfaces but significantly decreased in the cationic surface adherent monocytes/macrophages. Conversely, interleukin-8 (IL-8) expression was significantly decreased in cells adherent to the hydrophilic and anionic surfaces. Further analysis revealed that the hydrophilic and anionic surfaces inhibited monocyte adhesion and IL-4-mediated macrophage fusion into foreign body giant cells (FBGCs). Therefore, hydrophilic and anionic surfaces promote an anti-inflammatory type of response by dictating selective cytokine production by biomaterial adherent monocytes and macrophages. These studies contribute information necessary to enhance our understanding of biocompatibility to be used to improve the in vivo lifetime of implanted medical devices and prostheses.

Lipid-based microtubular drug delivery vehicles.

Lipid microtubules that self-assemble from a diacetylenic lipid are suitable structures for the sustained release of bioactive agents. Microtubules were loaded with agents under aqueous conditions and embedded in an agarose hydrogel for localization at areas of interest. Protein release from our microtubule-hydrogel delivery system was characterized in vitro, and in vivo biocompatibility was examined. The influences of protein molecular weight and initial loading concentration on release profile were evaluated by releasing test proteins myoglobin, albumin, and thyroglobulin. Protein molecular weight inversely affected the release rate, and loading with a higher protein concentration increased the mass but not the percent of initially loaded protein released daily. Preservation of protein activity was demonstrated by the ability of a neurotrophic factor released from the delivery system to induce neurite extension in PC12 cells. Bovine aortic smooth muscle cells co-cultured with the microtubule-hydrogel system showed no evidence of cytotoxicity and proliferated in the presence of the microtubules. Subcutaneous implantation of microtubules in rodents revealed no significant inflammatory response after 10 days. Our microtubule-hydrogel system is useful for applications where sustained release without contact between agent and organic solvents is desired.

eNOS-overexpressing endothelial cells inhibit platelet aggregation and smooth muscle cell proliferation in vitro.

Endothelial cell seeding of synthetic small diameter vascular grafts (SSDVG) has been shown to diminish thrombosis and intimal hyperplasia, resulting in improved graft patency. However, endothelial cell retention on seeded grafts when exposed to physiological shearing conditions remains poor. We report that the genetic engineering of endothelial cells to overexpress endothelial nitric oxide synthase (eNOS), may create improved anti-thrombotic and anti-hyperplastic endothelial cell phenotypes for SSDVG seeding. eNOS-overexpressing endothelial cells may potentially overcome the biochemical loss due to shear induced reduction in endothelial cell coverage on SSDVG. Bovine aortic endothelial cells (BAEC) were transfected with the human eNOS gene, and co-incubated with either human whole blood or bovine aortic smooth muscle cells (BASMC) in vitro. eNOS-transfected BAEC significantly overexpressed eNOS compared to control beta-Gal-transfected and untransfected BAEC up to 120 h post transfection. In co-incubation and co-culture assays, human platelet aggregation decreased by 46% and BASMC proliferation decreased by 67.2% when compared to incubation with untransfected BAEC.

High molecular weight kininogen inhibition of endothelial cell function on biomaterials.

Synthetic vascular grafts implanted into humans fail to develop a complete endothelial lining. In previous studies, we have shown that high-molecular-weight kininogens (HMWK) adsorb to the surfaces of biomaterials. In addition, it has been demonstrated that these proteins modulate cellular function. In the present study, we report on the adhesion and proliferation of human umbilical-vein endothelial cells (HUVEC) on tissue culture polystyrene, glass, polyurethane, and Mylar(trade mark) surfaces coated with human HMWK, either single-chain HMWK (SC-HMWK) or double-chain HMWK (DC-HMWK). Surfaces coated with fibronectin served as a positive control for these experiments. Parallel experiments were performed in which HUVEC were allowed to migrate from crosslinked dextran microcarrier beads (Cytodex 2) onto HMWK-coated surfaces. Our results indicate that HMWK-coated surfaces inhibit endothelial cell adhesion, proliferation, and migration at 24 and 72 h, and this inhibition is concentration dependent. To determine a potential mechanism for this inhibitory phenomenon, cells were stained for cytoskeletal actin filaments using rhodamine-phalloidin. Endothelial cells on HMWK-coated surfaces displayed F-actin filament reorganization/disassembly, characterized by the absence of peripheral actin bands in focal adhesion contacts. We conclude that HMWK inhibit endothelial cell adhesion, proliferation, and migration on a variety of biomaterial surfaces. This inhibitory effect may play a role in promoting the lack of endothelialization in synthetic vascular grafts, which is thought to play a significant role in the failure of these devices.

Cyclic-strain-induced endothelial cell expression of adhesion molecules and their roles in monocyte-endothelial interaction.

Vascular endothelial cells (ECs) are constantly subjected to hemodynamic forces that may regulate monocyte-endothelial interaction in vivo. To examine the effects of cyclic strain on endothelial expression of monocyte adhesion molecules, E-selectin, intercellular adhesion molecule-1 (ICAM-1), and vascular cell adhesion molecule-1 (VCAM-1) ECs were exposed to physiologically relevant levels of cyclic strain. When ECs were under 25% maximal strain at 30 cycles/min for 24 h, the expression of E-selectin significantly (p<0.05) increased, by 83%, compared to control ECs under static conditions. Similarly, monocyte adhesion to ECs under strain (maximum of 15 or 25% at 30 and 60 cycles/min for 24 h) also significantly (p<0.05) increased, by >82%. This cyclic-strain-induced monocyte adhesion was substantially inhibited (83.5%) by anti-E-selectin antibody. ICAM-1 expression also significantly increased, by 62%, when ECs were under 25% maximal strain at 30 cycles/min for 3 h whereas VCAM-1 expression by ECs under strain (for 0.5, 3, and 24 h) did not change compared to static ECs. When ECs were treated with anti-ICAM-1 antibody and monocytes with anti-VLA-4 antibody, an increase in monocyte adhesion to ECs under cyclic strain was reduced significantly. These results demonstrate that cyclic strain can induce EC expression of monocyte adhesion molecules E-selectin, ICAM-1, and VCAM-1 in a time-dependent manner and thus can mediate monocyte adhesion.

Cyclic strain effects on human monocyte interactions with endothelial cells and extracellular matrix proteins.

Human vascular endothelial cells (ECs) are exposed to various levels of hemodynamic forces, cyclic strain, and shear stress in vivo. Here, we examined the in vitro effects of the various levels (0-6%, 7-16%, and 17-25%) of strain at 60, 30, and 15 cycles per minute (cpm) on human monocyte adherence to endothelial cells and extracellular matrix protein preabsorbed surfaces. Monocyte adhesion to endothelial cells under cyclic strain significantly increased. At both 30 and 60 cpm, ECs under strains of 7-16% and 17-25% showed >52% and >117% higher monocyte adhesion than endothelial cells under static condition when monocytes were added for 0.5 h. This increase in monocyte adhesion to ECs under cyclic strain remained significantly higher even after 24 h of incubation. Human monocyte adhesion to extracellular matrix protein preabsorbed surfaces differed depending on the specific extracellular matrix protein. Monocytes adhered to collagen type I and fibronectin preabsorbed surfaces >50% under 0-6% strain, >23% under 7-16% strain, and >52% under 17-25% strain at 15 and 30 cpm compared to the collagen type V preabsorbed surface. However, when extracellular protein preabsorbed surfaces under cyclic strain were compared to the control static condition, monocyte adhesion did not significantly change on most of other surfaces. These results suggest that cyclic strain may play a role in the regulation of monocyte-endothelial cells/extracellular matrix interactions in vivo.

Estrogen modulates paracellular permeability of human endothelial cells by eNOS- and iNOS-related mechanisms.

Estradiol had a biphasic effect on permeability across cultures of human umbilical vein endothelial cells (HUVEC): at nanomolar concentrations it decreased the HUVEC culture permeability, but at micromolar concentrations it increased the permeability. The objective of the present study was to test the hypothesis that the changes in permeability were mediated by nitric oxide (NO)-related mechanisms. The results revealed dual modulation of endothelial paracellular permeability by estrogen. 1) An endothelial NO synthase (eNOS)-, NO-, and cGMP-related, Ca2+-dependent decrease in permeability was activated by nanomolar concentrations of estradiol, resulting in enhanced Cl- influx, increased cell size, and increases in the resistance of the lateral intercellular space (RLIS) and in the resistance of the tight junctions (RTJ); these effects appeared to be limited by the ability of cells to generate cGMP in response to NO. 2) An inducible NO synthase (iNOS)- and NO-related, Ca2+-independent increase in permeability was activated by micromolar concentrations of estradiol, resulting in enhanced Cl- efflux, decreased cell size, and decreased RLIS and RTJ. We conclude that the net effect on transendothelial permeability across HUVEC depends on the relative contributions of each of these two systems to the total paracellular resistance.

Acquired increase in leucocyte binding by intestinal microvascular endothelium in inflammatory bowel disease.

BACKGROUND: Endothelial cells that line microvascular blood vessels have an important role in inflammation through their ability to bind and recruit circulating leucocytes. Endothelial cells from the intestines of patients with chronically inflamed Crohn's disease and ulcerative colitis--the two forms of inflammatory bowel disease--display an increased leucocyte-binding capacity in vitro. We investigated whether this enhanced leucocyte binding is a primary or an acquired defect. METHODS: We cultured human intestinal microvascular endothelial cells (HIMEC) from the uninvolved intestine and chronically inflamed bowel of three patients with inflammatory bowel disease (two Crohn's disease, one ulcerative colitis). We assessed HIMEC binding to polymorphonuclear leucocytes and U937 cells by means of an adhesion assay. FINDINGS: After activation with interleukin-1beta or lipopolysaccharide, HIMEC from the chronically inflamed tissue in all three patients with inflammatory bowel disease bound twice as many polymorphonuclear leucocytes and U937 cells as endothelial cells from uninvolved tissue. INTERPRETATION: Enhanced leucocyte binding by HIMEC from chronically inflamed tissue in patients with inflammatory bowel disease is an acquired defect since it is not found in the uninvolved intestinal segments from the same individuals. Because interaction between endothelial cells and leucocytes is a key regulatory step in the inflammatory process, this enhanced binding may contribute to the pathophysiology of chronic intestinal inflammation.

Effects of estrogen on tight junctional resistance in cultured human umbilical vein endothelial cells.

OBJECTIVE: To study the effects of estrogen on transendothelial paracellular permeability in women. METHODS: Human umbilical vein endothelial cells (HUVEC) obtained from women were grown on filters. The paracellular permeability characteristics were determined in terms of changes in the permeability to the polar acid pyranine (Ppyr) and as changes in the transendothelial electrical resistance (RTE). Tight junctional resistance characteristics were assayed by lowering luminal NaCl and measuring the dilution potential, and were expressed as the ratio of monoion mobility uCl/uNa (cation selectivity). RESULTS: Low extracellular calcium and hyperosmolarity increased Ppyr and decreased RTE. The former but not the latter condition abolished the endothelium-specific cation selectivity. Treatment with 10 nM of estradiol-17 beta had no effect on RTE, but it increased the cation selectivity. The effect of estradiol required 1-6 hours' incubation with the hormone; it was dose dependent and saturable, with a median effective concentration of estradiol of 1 nM. Diethylstilbestrol, but not estriol, could mimic the effect of estradiol, and the estrogen receptor antagonist ICI-182, 780 blocked it. CONCLUSION: Cultured HUVEC cells form patent tight junctions. Estrogens increase the cation selectivity across HUVEC cultures. The effect of estrogen may be mediated by an estrogen receptor. These effects may be important for vasculoprotection in cases of sudden changes in ions levels across the capillary wall, such as ischemia or reperfusion.

iNOS expression in human intestinal microvascular endothelial cells inhibits leukocyte adhesion.

Increased nitric oxide (NO) production by inducible nitric oxide synthase (iNOS) has been associated with intestinal inflammation, including human inflammatory bowel disease. However, NO can downregulate endothelial activation and leukocyte adhesion, critical steps in the inflammatory response. Using primary cultures of human intestinal microvascular endothelial cells (HIMEC), we determined the role of NO in the regulation of HIMEC activation and interaction with leukocytes. Both nonselective (NG-monomethyl-L-arginine) and specific (N-iminoethyl-L-lysine) competitive inhibitors of iNOS significantly increased binding of leukocytes by HIMEC activated with cytokines and lipopolysaccharide. Increased adhesion was reversible with the NOS substrate L-arginine and was not observed in human umbilical vein endothelial cells (HUVEC). Activation of HIMEC significantly upregulated HIMEC iNOS expression and NO production. NOS inhibitors did not augment cell adhesion molecule levels in activated HIMEC but did result in sustained increases in intracellular reactive oxygen species. In addition, antioxidant compounds reversed the effect of NOS inhibitors on HIMEC-leukocyte interaction. Taken together, these data suggest that after HIMEC activation, iNOS-derived NO is an endogenous antioxidant, downregulating leukocyte binding and potentially downregulating intestinal inflammation.

Enhanced leukocyte binding by intestinal microvascular endothelial cells in inflammatory bowel disease.

BACKGROUND & AIMS: Microvascular endothelial cells mediate leukocyte homing, angiogenesis, and inflammation and healing and show tissue-specific adhesion molecules and functions. The activation of human intestinal mucosal microvascular endothelial cells (HIMECs) was studied in vitro to uncover possible abnormalities associated with inflammatory bowel disease. METHODS: HIMECs were isolated from normal and inflammatory bowel disease mucosa and assessed for phenotypic and morphological features, proliferative response, leukocyte binding capacity, and adhesion molecule expression. RESULTS: Basal proliferation by HIMECs was less than that of human umbilical vein endothelial cells (HUVECs) but increased proportionally more in response to vascular endothelial growth factor. Proinflammatory stimuli induced an activated, spindle-shaped morphology in HIMEC monolayers. Compared with HUVECs, unstimulated HIMECs showed less adhesiveness for U937 and MOLT4 cells and neutrophils, but cytokines and lipopolysaccharide substantially increased the binding capacity of HIMECs. HIMECs derived from inflammatory bowel disease mucosa showed a markedly greater leukocyte-binding capacity than normal mucosal HIMECs. Patterns of intercellular adhesion molecule 1, vascular cell adhesion molecule 1 and E-selectin messenger RNA expression were distinct in HIMECs, HUVECs, and mucosal mesenchymal cells. CONCLUSIONS: HIMECs represent differentiated endothelial cells with unique functional properties. Their dramatically enhanced capacity to bind leukocytes in inflammatory bowel disease suggests that HIMECs play an important role in initiating or maintaining inflammation.

Inhibition of migration of endothelial cells by type V collagen and high molecular weight kininogens.

The failure of synthetic vascular grafts in humans is complex and related not only to the type of material but also to the accumulation of cellular and noncellular components onto or within the material. Proteins that adsorb on biomaterials such as vascular grafts may be involved in modulating cell function. Of these components, type V collagen and high molecular weight kininogens (HMWKs) are known to localize on the surface of vascular grafts. We have investigated the effect of type V collagen and HMWKs in the prevention of endothelialization of vascular prostheses by inhibiting cell migration. Inhibition of endothelial cell migration may lead to incomplete endothelialization and thus, failure of these devices. Various concentrations of type V collagen and HMWK solutions were placed at the bottom of 24 well tissue culture plates with porous membrane inserts (8 microns pore size) placed on top. Suspended human umbilical vein endothelial cells (HUVECs) were placed on top of the insert and incubated at 37 degrees C for 2 hours. Following incubation, the inserts were fixed and the cells that migrated through the pores were counted by microscopy. Our results showed that migration of endothelial cells was significantly inhibited in a dose dependent manner by type V collagen and HMWKs.

Preparation of purified atactic polypropylene and polyvinyl methyl ether surfaces for thrombogenicity studies.

Commercial samples of atactic polypropylene (aPP) and polyvinyl methyl ether (PVME) were purified and spin-cast onto glass coverslips with a view to using these as model surfaces in thrombogenicity studies. These materials differ from polyvinyl alcohol (PVA) in a single functional group and are similarly amorphous: with the same backbone they have a hydroxyl, a methoxy, or a methyl group. The objective was to understand the role of the hydroxyl group in the platelet reactivity of PVA. Surface characterization showed that they were chemically pure (as determined by X-ray photoelectron spectroscopy) but not smooth (as determined by scanning electron microscopy or interferometry), presumably due to the difficulties of spin-casting optically clear films from hot solutions (aPP or polyethylene [PE]) or because of imperfect adhesion to the saline-treated substrate (PVME). PVME was also gamma-irradiated to insolubilize it. Fewer platelets adhered to PVA than to PVME or to aPP and PE, but roughness effects and limited data preclude definitive conclusions regarding the effect of functional groups. Less protein was found on PVA than on the hydrophobic surfaces, but the significance of this observation is unclear. Further studies with more sensitive protocols are called for to examine the extent of platelet activation and its relationship to surface chemistry.

Surface-bound heparin fails to reduce thrombin formation during clinical cardiopulmonary bypass.

The hypothesis that heparin-coated perfusion circuits reduce thrombin formation and activity; fibrinolysis; and platelet, complement, and neutrophil activation was tested in 20 consecutive, randomized adults who had cardiopulmonary bypass. Twenty identical perfusion systems were used; in 10, all blood-contacting surfaces were coated with partially degraded heparin (Carmeda process; Medtronic Cardiopulmonary, Anaheim, Calif.). All patients received a 300 U/kg dose of heparin. Activated clotting times were maintained longer than 400 seconds. Cardiopulmonary bypass lasted 36 to 244 minutes. Blood samples for platelet count, platelet response to adenosine diphosphate, plasma beta-thromboglobulin, inactivated complement 3b, neutrophil elastase, fibrinopeptide A, prothrombin fragment F1.2, thrombin-antithrombin complex, tissue plasminogen activator, plasminogen activator inhibitor-1, plasmin alpha 2-antiplasmin complex, and D-dimer were obtained at these times: after heparin was given, 5 and 30 minutes after cardiopulmonary bypass was started, within 5 minutes after bypass was stopped, and 15 minutes after protamine was given. After cardiopulmonary bypass, tubing segments were analyzed for surface-adsorbed anti-thrombin, fibrinogen, factor XII, and von Willebrand factor by radioimmunoassay. Heparin-coated circuits significantly (p < 0.001) reduced platelet adhesion and maintained platelet sensitivity to adenosine diphosphate (p = 0.015), but did not reduce release of beta-thromboglobulin. There were no significant differences between groups at any time for fibrinopeptide A, prothrombin fragment F1.2, or thrombin-antithrombin complex or in the markers for fibrinolysis: D-dimer, tissue plasminogen activator, plasminogen activator inhibitor-1, and alpha 2-antiplasmin complex. In both groups, concentrations of prothrombin fragment F1.2 and thrombin-antithrombin complex increased progressively and significantly during cardiopulmonary bypass and after protamine was given. Concentrations of D-dimer, alpha 2-antiplasmin complex, and plasminogen activator inhibitor-1 also increased significantly during bypass in both groups. Fibrinopeptide A levels did not increase during bypass but in both groups increased significantly after protamine was given. No significant differences were observed between groups for levels of inactivated complement 3b or neutrophil elastase. Radioimmunoassay showed a significant increase in surface-adsorbed antithrombin on coated circuits but no significant differences between groups for other proteins. We conclude that heparin-coated circuits used with standard doses of systemic heparin reduce platelet adhesion and improve platelet function but do not produce a meaningful anticoagulant effect during clinical cardiopulmonary bypass. The data do not support the practice of reducing systemic heparin doses during cardiac operations with heparin-coated extracorporeal perfusion circuitry.

Bacteria/blood/material interactions. I. Injected and preseeded slime-forming Staphylococcus epidermidis in flowing blood with biomaterials.

Blood-material interactions were studied using in vitro recirculation with human blood, slime-forming Staphylococcus epidermidis, and cardiovascular materials. Staphylococcus epidermidis, under preseeded or injected conditions, adhered to nonsmooth materials and elevated plasma levels of fibrinopeptide A (FpA) and C3a in the presence of all materials. Increased white blood cell (WBC) and platelet adhesion and thrombospondin and platelet factor 4 (PF4) release were noted for respective materials in the presence of injected bacteria. Materials that adhered significant quantities of injected S. epidermidis exhibited low levels of adsorbed proteins. Materials with high levels of preseeded S. epidermidis showed high levels of adsorbed proteins. Adhesion of preseeded bacteria and blood plasma elevations of C3a and FpA were lowest on semicrystalline polymer substrates, intermediate on halogenated substrates, and highest on amorphous substrates. In the presence of injected bacteria, WBCs and platelets adhered at earlier recirculation times to amorphous substrates than to semicrystalline substrates.

Protein adsorption and macrophage activation on polydimethylsiloxane and silicone rubber.

Static and dynamic human blood adsorption studies on polydimethylsiloxane, PDMS, and silicone rubber show that these materials are similar, but not identical, in their protein adsorption behavior. Fibrinogen, immunoglobulin G, and albumin were the predominant proteins identified on the material surfaces with fibronectin, Hageman factor (factor XII), and factor VIII/vWF adsorbing at intermediate levels. While the protein adsorption characteristics for the two materials were similar, higher levels of the respective proteins were identified on silicone rubber compared to PDMS. Monocytes/macrophages incubated on PDMS, silicone rubber and low density polyethylene, LDPE, with or without protein adsorption produced variable levels of IL-1 beta, IL-6 and TNF-alpha dependent on the polymer and adsorbed protein. PDMS showed lower levels of the cytokines when compared to the polystyrene control and polyethylene. Protein preadsorption on the PDMS, polystyrene, and LDPE surfaces showed lower levels of cytokines when compared to the respective quantities produced with no protein adsorption suggesting a passivating effect by the protein adsorption phenomenon on monocyte/macrophage activation. Preadsorption of IgG, fibrinogen or fibronectin decreased the quantitative expression of IL-1 beta but increased the functional activity in the thymocyte proliferation assay indicating the presence of monocyte/macrophage activation products which either downregulated the activity of IL-1 beta or upregulated thymocyte proliferation in an independent fashion.

Problems and approaches in covalent attachment of peptides and proteins to inorganic surfaces for biosensor applications.

Some of the fundamental problems in covalent attachment of peptides and proteins to putative biosensor surfaces are reviewed and specific approaches to these problems discussed. In addition, selected aspects of our recent work utilizing self-assembled monolayer (SAM) systems designed to react selectively with the thiol side chain of Cys in proteins are presented. Uniform attachment of a 21-amino acid peptide antigen through a single Cys residue with retention of biological function (antibody binding) has been attained. Further work with this system may lead to solutions for some of the problems which currently prevent the development of reliable biosensors for industrial and medical use.

Immobilization of high-affinity heparin oligosaccharides to radiofrequency plasma-modified polyethylene.

Oligosaccharides of heparin with high affinity for antithrombin III (ATIII) have been immobilized onto surface-modified NHLBI Primary Reference low density polyethylene (PE). PE was modified by radiofrequency plasma polymerized (< 150 nm thick) films derived from N-vinyl-2-pyrrolidone (PPNVP) or allyl alcohol (PPAA), and coupled by chemical derivatization to either 3-aminopropyltriethoxysilane or amino-terminated poly(ethylene oxide). High affinity heparin oligosaccharides (HA-heparin, anti-factor Xa activity of 592 +/- 120 IU/mg) prepared by partial deaminative cleavage of commercial crude heparin and fractionated by agarose-ATIII affinity chromatography, were immobilized to surface-modified PE by reductive amination. The anticoagulant activity, as determined by a chromogenic assay for the inhibition of factor Xa, was estimated to be 30-70 mIU/cm2, with binding estimated to be 56-119 ng/cm2. The highest activity was obtained for the HA-heparin immobilized to PE modified by PPNVP with a PEO spacer. Visual confirmation of ATIII binding to immobilized HA-heparin was demonstrated by a gold-labeled double antibody method with imaging by SEM.

Human vascular endothelial cell attachment and growth inhibition by type V collagen.

PURPOSE: Human vascular prostheses develop a pseudointima that after time is devoid of a complete endothelial lining. The composition of this pseudointima consists of both cellular and noncellular components such as coagulation proteins and extracellular matrix proteins. Of these extracellular molecules, type V collagen has been reported to be localized to surfaces of vascular prostheses. We have hypothesized that type V collagen may be involved in the mechanisms of inhibition of endothelialization on vascular prostheses. In this study, interactions of human protein and cells with clinically used material and reference surfaces were analyzed in vitro. METHODS AND RESULTS: Human collagen types IV and V or human fibronectin was coated on disks punched from expanded polytetrafluoroethylene, bacteriologic polystyrene, and tissue culture-treated polystyrene. Fibronectin adsorbed equally to these surfaces, but differential adsorption of type V collagen occurred. The attachment and growth of human saphenous vein and umbilical vein endothelial cells and of human skin fibroblasts were also evaluated on protein-coated or uncoated surfaces. Type IV collagen and fibronectin promoted the attachment of these cells, but type V collagen reduced cellular adhesion. Growth of endothelial cells was significantly inhibited on surfaces coated with type V collagen even when additional growth substances such as serum, retinal-derived growth factor, and heparin were present in the medium. Human adult dermal fibroblast adhesion and cell growth were not affected by coating the surfaces with type V collagen. CONCLUSIONS: The components of the extracellular matrix of the pseudointima may directly influence endothelial cells by inhibiting cell proliferation, migration from within the graft or from anastomoses, or both. Type V collagen, a matrix protein found at luminal surfaces of vascular prostheses, may be one protein responsible for control of endothelial responses.