In Vitro and In Silico Studies of Functionalized Polyurethane Surfaces toward Understanding Biologically Relevant Interactions

The solid–aqueous boundary formed upon biomaterial implantation provides a playground for most biochemical reactions and physiological processes involved in implant–host interactions. Therefore, for biomaterial development, optimization, and application, it is essential to understand the biomaterial–water interface in depth. In this study, oxygen plasma-functionalized polyurethane surfaces that can be successfully utilized in contact with the tissue of the respiratory system were prepared and investigated. Through experiments, the influence of plasma treatment on the physicochemical properties of polyurethane was investigated by atomic force microscopy, attenuated total reflection infrared spectroscopy, differential thermal analysis, X-ray photoelectron spectroscopy, secondary ion mass spectrometry, and contact angle measurements, supplemented with biological tests using the A549 cell line and two bacteria strains (Staphylococcus aureus and Pseudomonas aeruginosa). The molecular interpretation of the experimental findings was achieved by molecular dynamics simulations employing newly developed, fully atomistic models of unmodified and plasma-functionalized polyurethane materials to characterize the polyurethane–water interfaces at the nanoscale in detail. The experimentally obtained polar and dispersive surface free energies were consistent with the calculated free energies, verifying the adequacy of the developed models. A 20% substitution of the polymeric chain termini by their oxidized variants was observed in the experimentally obtained plasma-modified polyurethane surface, indicating the surface saturation with oxygen–containing functional groups.


Comparison of the developed in silico model with previous computational models of polyurethane
To compare the developed in silico polyurethane model with the models published previously [Ref. 24,25], we analyzed the structural properties of bulk polyurethane obtained by utilizing our force field with those available in the literature.To this end, we calculated pair radial distribution functions (RDF) between the individual fragments in our model (FRA, FRB, FRC) and compared them with the previously published data.A direct comparison is complicated because of differences in definitions of groups along the polymer chain.Also, the simulated polyurethane materials have different characteristics (length of chains, initial conditions etc.) The RDF calculated between the elastic FRB-FRB fragments (depicted in Fig. S2., blue) shows a well-pronounced peak at r=~0.5 nm with the value of ~2, followed by a less-pronounced peak at ~0.9 nm, and a weak peak at ~1.3 nm.Such behavior of RDF arises from the liquid-like structure of the chain fragments of the material, and similar behavior of a corresponding RDF was observed in Ref. 25.It also agrees with the disordered structure observed for our material via XRD (Fig. S3).The RDF calculated between the elastic and non-elastic fragments (FRA-FRB and FRC-FRB) shows no clear maxima, reporting on rather repulsive interactions between these fragments.Again, it demonstrates that the bulk of the simulated material is rather unstructured.Similar behavior for elastic-nonelastic fragments was observed in both previous models Ref. [24,25].In summary, the computationally-extensive model developed here shows overall agreement regarding structural data with previous computational models of polyurethane materials.

Fig. S3 .
Fig. S3.The XRD results of studied polyurethane.The XRD data were collected on a Rigaku Miniflex System, Cu Kα radiations at 10 mA and 10 kV, with the diffraction scanning rate of 1°/min.

Fig. S4 .
Fig. S4.Radial distribution functions calculated between oxygen atoms in the polyurethane fragments and the oxygen atom of water of the not oxidized system.Both interfacial and bulk groups are considered as most of water-unoxidized polymer contacts occur at the interface.The functions are not normalized to unity because of the asymmetry of the interface.The analysis demonstrates that most of water-unoxidized polymer contacts occur at FRC fragments.

Fig. S5 .
Fig. S5.Radial distribution functions calculated between oxygen atoms in the polyurethane fragments and the oxygen atom of water of the oxidized system.For COX1, also the function for hydroxyl oxygen (Oh)-water oxygen is shown to demonstrate that most of water-oxidized polymer interactions occur between hydroxyl groups of oxidized fragments and water.Only sub-interfacial groups (>1 nm from the interface) are considered to analyze the penetration of the oxidized polymer by water.The functions are not normalized to unity because of the asymmetry of the interface.

Fig. S6 .
Fig. S6.Number density (atoms per nm 3 ) profiles of oxygen atoms in the polyurethane structure and water oxygen calculated for unoxidized and oxidized systems.For the polymer, full profiles along the MD simulation box are shown with zero at the y-axis denoting the midplane of the polymer placed in the middle of the simulation box.Oxidation changes the distribution of oxygen at the boundaries.In the sub-interfacial region, the oxygen density does not significantly change after oxidation because an increase due to oxidation is compensated by the reduced overall number of oxygen atoms in the truncated oxidized fragments.