A major area Ofbiomaterials research is the development of surfaces that reduce or eliminate non-specific protein adsorption. End-tethered PEO has been shown to reduce protein and cell interactions at the tissue-material interface;​ the effects of polymer chain length, chain density and end-group chemistry are not yet completely understood. To date, there have been few detailed, systematic studies that have attempted to elucidate the effect of end-tethered PEO conformation, surface chain density, molecular weight (MW) and end-group chemistry on protein adsorption at the solid-liquid interface.

In the research described in this thesis PEOs of varying molecular weight (600, 750, 2000 and 5000 MW) and terminal functional group (-OH, -OCH₃) were thiolated and chemisorbed to gold coated silicon wafers for the purpose of characterizing film thickness and surface chain density for direct correlation to protein adsorption behaviour. Tethered chain density was varied by manipulating PEO solubility and chemisorption time, which in principle, should allow for variable, controlled surface chain density from low to very high values. PEO layers were characterized using water contact angles, X-ray photoelectron spectroscopy (XPS), self-nulling ellipsometry and neutron reflectometry (NR). The adsorption of two proteins having widely different molecular weights was examined using radiolabeling and ellipsometry to ascertain the effectiveness of these surfaces in resisting protein adsorption and to provide information about the nature of protein interactions with end-tethered PEO surfaces. These experiments were carried out using single or binary protein solutions in buffer. Adsorption from plasma was also investigated:​ (1) by Western Blot analysis ofthe proteins eluted after plasma contact;​ (2) via experiments using radiolabeled fibrinogen.

The chemisorption ofthiolated PEO to gold coated silicon wafers proved to be an effective method for producing surfaces with variable, controlled chain densities. Also, it was apparent that the chain densities obtained were not only among the highest ever reported for these systems, but also that the range of achievable chain density was very broad for all molecular weights studied. Neutron reflectometry and ellipsometry measurements gave values of chain density that were the same within experimental error. For the 750 MW PEO, in situ neutron reflectometry yielded novel information about the evolution of layer structure, as the surface filled, for layers formed under high and low solubility conditions. It was concluded from these experiments:​ (1) that protein adsorption passed through a minimum as PEO volume fraction increased, the minimum occurring at a PEO volume fraction of ~0.39. This is a novel observation for PEO-based inhibition of protein adsorption;​ (2) that at high chain density there may be regions within the PEO layer where the effective concentration is above the solubility limit of PEO in aqueous solution. It is hypothesized that in these regions the polymer has been “forced” from solution, forming hydrophobic patches that facilitate increased protein adsorption at high chain density.

Initial single protein adsorption experiments using radiolabeled proteins showed that resistance to fibrinogen passed through a maximum as chain density increased and that at a chain density of ~0.5 chains∕nm² protein adsorption was suppressed to about the same extent (80% decrease in adsorption compared to unmodified gold) for 750 and 2000 MW PEO layers. Further study of single protein adsorption using in situ ellipsometry showed similar trends:​ i.e. adsorption minima for the 750 and 2000 MW systems were similar and occurred at similar chain densities (in this case for both fibrinogen and lysozyme). The observed fibrinogen and lysozyme adsorption trends suggest that chain density, not chain length, is the major determinant of protein resistance. Furthermore, at chain densities of ~ 0.5 chains∕nm², protein resistance (expressed as fractional reduction compared to the control) seems to be independent of protein size. It is hypothesized that at high chain density, the chemisorbed PEO becomes dehydrated yielding a surface that is less protein resistant. This idea is substantiated by neutron reflectometry data.

Experiments on adsorption from plasma were used to evaluate the surfaces determined to be ‘optimal’ in reducing single protein adsorption. Vroman effect type experiments, using ¹²⁵I labeled fibrinogen added to the plasma as a tracer, showed that PEO layers formed from solutions near the cloud point adsorbed the lowest amounts of fibrinogen. Layers of OH-terminated 600 MW PEO showed almost complete suppression (versus controls) of the Vroman peak. Respective Vroman peak amounts of adsorbed fibrinogen for 600-OH, 750-OCH₃ and 2000-ΘCH₃ were 20 ± 1, 70 ± 20, and 50 ± 3 ng∕cm² compared to 400 ng∕cm² for unmodified gold;​ adsorption levels at higher plasma concentration were 6.7 ± 0.6, 16 ÷ 9 and 12 ± 3 ng∕cm² respectively compared to 150 ng∕cm² for gold. Fibrinogen adsorption from plasma was not significantly different for surfaces prepared with PEO of molecular weight 750 and 2000 when the chain density was the same (~0.5 chains∕nm²) supporting the conclusion that chain density may be the key property for suppression of protein adsorption. Furthermore, the data suggested a distal chain end group effect such that fibrinogen adsorption was reduced and/or its displacement from the surface facilitated, on hydroxyl-terminated compared to methoxyterminated PEO layers.

SDS-PAGE gels and immunoblots ofthe proteins eluted from these surfaces after plasma contact showed that a number of proteins were adsorbed, including fibrinogen, albumin, C3 and apolipoprotein A-L However, the blot responses were weak for all four proteins of the contact system;​ some complement activation was observed on all of the surfaces studied.