An image analysis strategy to study cell adhesion

A semi-automated in situ technique has been developed for the study of the extent and kinetics of cell adhesion at the individual cell level. Our investigation involves the static sedimentation of glutaraldehyde-fixed human erythrocytes suspended in 10 mM NaCl or 10 mM NaCl containing 2% (v/v) 1-propanol onto flat, horizontal, and transparent surfaces. The surfaces used are glass, poly(ethylene terephthalate), polystyrene, and fluorinated ethylene propylene. An inverted microscope is utilized for observations. Brownian motion is used as the distinguishing criterion between adherent and non-adherent cells. The extent of adhesion is expressed as the percentage of adherent cells. Two digital image processing techniques, image averaging and image subtraction, are presented for automation of the methodology. Although all non-adherent cells undergo Brownian motion, they exhibit this behavior to varying degrees. Factors under consideration are the liquid medium's surface tension (γ_LV) and the solid substrate surface tension (γ_SV). Preliminary results reveal that, in general, variations of γ_SV and γ_LV have a statistically significant effect on the extent of adhesion at the 99% and 96% confidence levels, respectively. A time depepdence for the adhesion of populations of cells is observed. However, individual cells either instantly or gradually adhere. Image subtraction generally overestimates the number of adherent cells due to the difficulty in detection of minute oscillations. The deviation between the adhesion percentage obtained from visual observations of the monitor and image subtraction is less than 10%.