Interpretation of cell mechanical experiments in microfluidic systems depend on the choice of cellular shape descriptors

Abstract

The capability to parameterize shapes is of essential importance in biomechanics to identify cells, to track their motion, and to quantify deformation. While various shape descriptors have already been investigated to study the morphology and migration of adherent cells, little is known of how the mathematical definition of a contour impacts the outcome of rheological experiments on cells in suspension. In microfluidic systems, hydrodynamic stress distributions induce time-dependent cell deformation that needs to be quantified to determine viscoelastic properties. Here, we compared nine different shape descriptors to characterize the deformation of suspended cells in an extensional as well as shear flow using dynamic real-time deformability cytometry. While stress relaxation depends on the amplitude and duration of stress, our results demonstrate that steady-state deformation can be predicted from single cell traces even for translocation times shorter than their characteristic time. Implementing an analytical simulation, performing experiments, and testing various data analysis strategies, we compared single cell and ensemble studies to address the question of computational costs vs experimental accuracy. Results indicate that high-throughput viscoelastic measurements of cells in suspension can be performed on an ensemble scale as long as the characteristic time matches the dimensions of the microfluidic system. Finally, we introduced a score to evaluate the shape descriptor-dependent effect size for cell deformation after cytoskeletal modifications. We provide evidence that single cell analysis in an extensional flow provides the highest sensitivity independent of shape parametrization, while inverse Haralick's circularity is mostly applicable to study cells in shear flow.