Biofilms represent bacterial communities that attach to implanted medical devices such as sutures and catheters. When detached from implants within the body, biofilms can cause chronic infections. This study used Simpleware software to generate image-based microscopic Finite Element (FE) models from confocal laser scanning microscopy scans of real biofilm colonies attached to a surgical suture fragment. The goal of the study was to predict how suture deformation affects the behavior of biofilm colonies. The FE method was used to solve continuum mechanics equations for various loading cases, enabling the simulation of the effect of tension, torsion and bending of a biomaterial structure on biofilm movement and detachment.
Engineering and Physical Sciences Research Council (EPSRC) (Grant EP/F034296/1); University of Southampton
Biofilms were grown from Straphylococcus aureus, a clinical methicillin resistant (MRSA) strain of bacteria. Red and black colonies were selected and re-streaked out until a single black colony phenotype remained. Pieces of size 4 Ti-Cron™ braided polyester were cut into ~2-cm lengths and combined with the bacteria.
The suture and the biofilm were then imaged using confocal laser scanning microscopy (CLSM), specifically a Leica DM RXE upright microscope attached to a TCS SP2 AOBS confocal system. 3D image stacks were collected using different zoom settings and Z-slices.
The CLSM images were loaded in Simpleware ScanIP as three image stacks, with each stack corresponding to a RGB channel containing 28 images. Semi-manual image segmentation involved extracting closed surfaces corresponding to the 3D boundaries of the suture and a series of disconnected 3D surfaces corresponding to the biofilm phase.
The 3D segmented surfaces were converted into a volume mesh in Simpleware FE, and were optimised to prevent the creation of an excessively high number of finite elements. An adaptive meshing technique was used to preserve surface topologies, and sub-regions were meshed as independent disconnected regions with node and face-matching contact interfaces for multi-part modelling.