Engineering Strategies for Sustainable Endothelialization of a New Type of Ventricular Assist Device

Abstract

Heart failure affects 1-2 % of the population in industrialized nations. Its progression, advanced heart failure is an acute, potentially fatal condition. If a patients heart pumping function is insufficient, their only option of survival becomes either one of few available heart transplants or a mechanical circulatory support system. The latter, termed Ventricular Assist Devices (VAD) are mobile electromechanical pumps that can compensate for insufficient heart pump function. Since their invention in the 1960 they have continuously evolved, now representing a viable permanent alternative for patients ineligible for heart transplantation. However, the contact of blood with the artificial device material neccessitates an aggressive anticoagulation regime to prevent thrombosis. This causes bleeding and stroke, two major complications that persist in current generations of these devices. Housed within the umbrella of the Zurich Heart Consortium, the hybrid membrane project envisions to solve this problem by proposing a new type of VAD: A pulsatile device akin to a diaphragm pump featuring an endothelial cell (EC) layer on the entire interior, providing optimal haemocompatiblity. The most demanding challenge of this project is the maintenance of a stable endothelium on core part of the device, the cyclically deforming elastomer membrane actuating the blood. The ECs will thus be subjected to superphysiological wall shear stress up to 10 Pa while simultaneously experiencing the substrate deformation of around 10 %. From this ambitious goal derive two key challenges which are addressed within this thesis: (1) the establishment of an in vitro laboratory test system capable of replicating the stated conditions and (2) an surface-based solution capable of sustaining the EC on the elastomer membrane in this environment. In collaboration with other groups from the hybrid membrane track we successfully established such a bioreactor system. We showed that we could image an endothelialized elastomer membrane and expose it to flow. A computational fluid dynamics simulation revealed that shear patterns which are highest for the region where the flow hits the membrane as it is inflated into the channel. Actuating the membrane at 1 Hz to an apex displacement of about 8 %, we found a good correlation for the alignment of the cells with the simulation as the maximum shear was just above 5 Pa. Raising the flow to generate values of 10 Pa of maximum shear led to the eroding of cells from this zone. These results serve as a good validation of our system and showed that we could replicate the mode of failure for the new VAD that needs to be overcome by tailored engineering solutions. We then designed and fabricated a honeycomb-like surface whose wells could shelter up to 2 ECs, were found to increase the cell density (by 20 %) and proliferation (by 40 %) and would reduce the WSS at the bottom of the well by up to 60 % (as shown by CFD simulations). We subsequently evaluated endothelialized substrates bearing this topography in both a flow only system featuring cyclic olefin copolymer as the substrate as well as in the above described dynamic bioreactor test system featuring PDMS as a substrate. We found that after exposure to 8 Pa and 10 Pa for 18 h there were above 40 % more cells on the structured substrates and also the connectivity was significantly better. We also found 90% and 40% reductions of the nuclear localizations of the inflammatory markers V-CAM and I-CAM when comparring structured samples to flat controls. For combined loading of about 10 Pa of apex WSS and 5 % and 10 % of WD we found density ratios of about 2 fold % for structured vs. unstructured membranes. These results demonstrate that the introduction of the honeycomb topography is very promising with respects to EC retention under superphysiological loading conditions. Our two achievements represent important contributions towards the development of the envisioned novel VAD and might potentially see clinical integration in the near future.