Characteristics of Cell Shape in Two Dimensions

Abstract

The shape of a cell is closely related with its properties, and in certain conditions it has been shown that change in cell shape can change cell behavior and outcome. However in two dimensions cells have complicated outlines, and it is not yet clear how best to characterize cell shape. Here we study the characteristics of cell shape of the 10T1/2 cell line in two-dimensional culture, and a few osteosarcoma cancer cell lines. The two dimensional surfaces are treated to make them either hydrophobic or hydrophilic, and the cells imaged after fixing, using membrane and actin labeling. The cells are also treated with pharmacological modulators of the cytoskeleton, yielding a large number of different shape types, and shape perturbations. Shape parameters are calculated by first using image processing to obtain binary outlines of the cells, and then calculating a number of geometric parameters, such as area, perimeter length, aspect ratio, fractal dimension etc. Simultaneously we also use orthogonal polynomial decompositions using Zernike moments to calculate shape characteristics. Using statistical data analysis we compare the ability of the Zernike moment expansion to capture the different shapes and their perturbations with that of the geometric parameters. We find that both types of shape calculations give insights into how a cell determines its shape on a surface.