Automated surface and volume measurement of fused cells

To design and evaluate an algorithm to automatically calculate membrane area, volume and derived values from grey scale microscopic images of pairs of spheroid fused cells, especially human erythrocytes.

Pairs of fused cells ("doublets") were identified by their high optical contrast, which resulted from their unhemolysed state. Global thresholding and noise-removing algorithms were applied to the image and resulted in a binary, "8"-shaped contour. The contour was used as a base for the calculation of a two-dimensional weighted distance histogram, the hilltops of which could be identified as center points of the contour's spheres. This allowed calculation of the distance of both center points and of the spheres' radii. With these three values, calculation of membrane area, volume and other derived values of the doublets became possible. High-speed time series were created based on consecutive images of the postfusion swelling and hemoglobin ejection from erythrocyte doublets at different temperatures.

The influence of observation temperature on the dynamics of electrofused erythrocytes was measured with the algorithms given, and results were in agreement with physical changes in cell plasma viscosity. Images taken by optical and electron microscopy were in agreement with the two-spheres model of a doublet. The algorithm was not affected by fragmented contours.

The velocity of hemoglobin ejection from electrofused erythrocytes and the relative change in static membrane area increase with temperature. The algorithm delivered reliable geometric values of fused cell configurations.