Abstract
Implantable vascular devices typically interface with blood and vascular tissues. Physical properties of device materials and coatings, independent of chemical composition, can significantly influence cell responses and implant success. Here, we analyzed the effect of various polymer processing regimes, using a single implant polymer - poly(ε-caprolactone) (PCL), on vascular endothelial cell (EC), smooth muscle cell (SMC), and platelet response. PCL films were formed by varying three parameters: 1) formation method - solvent casting, melt pressing or spin coating; 2) molecular weight - 50 or 100 kDa; and 3) solvent type - dichloromethane (DCM) or tetrahydrofuran (THF). We quantified the relationship of polymer processing choice to surface roughness, wettability, and bulk stiffness; and to EC adhesion, SMC adhesion, and platelet activity state (PAS). Multiple regression analysis identified which processing method signficantly impacted (F-ratio>p-value; p<0.1) polymer physical properties and vascular cell interaction. Film formation method affected PCL roughness (Rq), wettability (°), and stiffness (MPa) with spin coating resulting in the most wettable (81.8±0.7°), and stiffest (1.12±0.07 MPa; p<0.001) polymer film; however, solvent cast films were the roughest (281±66nm). Molecular weight influenced wettability, with the highest wettability on 50 kDa films (79.7±0.7°; p<0.001) and DCM solvent films (83.0±1.0°; p<0.01). The multiple regression model confidently predicted (F-ratio=9.88; p=0.005) wettability from molecular weight (p=0.002) and film formation method (p=0.03); stiffness (F-ratio=4.21; p=0.05) also fit well tofilm formation method (p=0.02). Film formation method impacted SMC adhesion and platelet activity state, but not EC adhesion, with melt press PCL promoting the highest SMC adhesion (18000±1536 SMCs; p<0.05) and PAS (5.0±0.7 %PAS). The regression model confidently fit SMC adhesion (F-ratio=3.15; p=0.09) and PAS (F-ratio=5.30; p=0.05) to polymer processing choices, specifically film formation method (p<0.03). However, only SMC adhesion had a model that fit well (F-ratio=4.13; p=0.05) to the physical properties directly, specifically roughness and wettability (p<0.04).