Advanced MD and CFD models for the design of magnetophoretic-microfluidic devices for blood detoxification processes

C. González-Fernández, J. Gómez-Pastora, A. Basauri, M. Fallanza, E. Bringas, J.J. Chalmers, I. Ortiz
University of Cantabria,
Spain

Keywords: ligand-receptor interactions, molecular dynamics simulations, particle magnetophoresis, CFD

Summary:

Magnetophoretic microseparator devices (MSDs) are being conceived as promising tools to be employed in blood detoxification processes, where dangerous pathogens get attached to functionalized magnetic beads and the resultant complexes are later magnetically recovered. For example, the application of MSDs for the direct removal of sepsis-causing agents, such as Lipid A, has been considered the most direct conceivable treatment against this infectious illness. The success of the process relies on both the capture of the pathogen with high affinity and selectivity and its later separation from blood by exerting a magnetic force on the particles. Nevertheless, the ligands that could be employed for the removal of this lipid or the optimization of the MSDs in order to process high flow rates are topics that, although they have not been thoroughly covered, are indispensable to bring this technology a step forward. Herein, we combined Molecular Dynamics (MD) and Computational Fluid Dynamics (CFD) simulations to address the design of such MSDs. Thereby, the MD software GROMOS11 and the GROMOS 54A8 force field were used to predict the most appropriate ligand by elucidating at atomic and molecular levels its interaction mechanism with Lipid A. In this work, we tested the anti-LPS factor rALF as a potential candidate, since it is an effector involved in the innate immune response against Lipid A infection of black tiger shrimp Penaeus monodon, and thus naturally interacts with Lipid A. The simulated conditions (300 K in a NaCl buffer and [NaCl]=150 mM) emulated the experimental conditions of the incubation step and the strength of the rALF-Lipid A binding was estimated by the Linear Interaction Energy method. Our results suggest that the Lipid A binds rALF with an affinity constant value of 1.10·103 M-1, which is similar to other potential ligands reported in the literature. Furthermore, the hydrophobic interactions between the acyl chains of Lipid A and the hydrophobic residues of rALF are key for the interaction. On the other hand, the magnetic separation was simulated with a customized CFD software, Flow-3D, which was used to predict the particle separation from blood streams in continuous-flow Y-Y shaped MSDs. In order to treat high blood flow rates while entirely collecting the particles and keeping unimpaired the blood quality, the channel length and cross section were first studied. We observed that working with rectangular, 10 mm long channels, the blood flow rate that can be treated could increase as much as 100 times in comparison to 2 mm long, U-shaped channels. Once the channel design was optimized, a parallelization process was later carried out to meet the flow rate requirements of typical detoxification systems. For that purpose, the magnets must be carefully selected and located so as to the magnetic force developed inside the channels is sufficient for capturing the beads at high flow rates. Overall, the computer-aided design guidelines reported in this work prove quite useful for addressing the optimization of MSDs, thus contributing to the feasible application of such lab-on-a-chip devices for blood detoxification processes.