Cell mechanics can be robustly derived from AFM indentation data using the brush model: error analysis

The brush model was introduced to interpret AFM indentation data collected on biological cells in a more consistent way compared just to the traditional Hertz model. It takes into account the presence of non-Hertzian deformation of the pericellular brush-like layer surrounding cells (a mix of glycocalyx molecules and microvilli/microridges). The model allows finding the effective Young's modulus of the cell body in a less depth-dependent manner. In addition, it allows finding the force due to the pericellular brush layer. Compared to simple mechanical models used to interpret the indentation experiments, the brush model has additional complexity. It raises the concern about the possible unambiguity of separation of mechanical properties of the cell body and pericellular layer. Here we present the analysis of the robustness of the brush model and demonstrate a weak dependence of the obtained results on the uncertainties within the model and experimental data. We critically analyzed the use of the brush model on a variety of AFM force curves collected on rather distinct cell types: human cervical epithelial cells, rat neurons, and zebrafish melanocytes. We conclude that the brush model is robust; the errors in the definition of the effective Young's modulus due to possible uncertainties of the model and experimental data are within 4%, which is less than the error, for example, due to a typical uncertainty in the spring constant of the AFM cantilever. We also discuss the errors of parameterization of the force due to the pericellular brush layer.