Similarity of sister-cell trajectories in fibroblast clones.

Abstract

ABSTRACTThe observation that sister cells of a fibroblast division execute somewhat symmetric trajectories has led to the controversial hypothesis that cell pathways on a substrate are preprogrammed in detail. We have investigated symmetry and structure in the paths of 170 human AG1519 sister-fibroblast pairs recorded on film and analysed by interactive computer techniques.In a set of plots of sister-cell trajectories recorded immediately after cell division, we found an average of 31 % mirror symmetry and 14 % rotation identity after rejecting tracks that crossed over each other, a result in substantial agreement with published data. Such ‘symmetries’ are quite inexact, however. A symmetry-controlled perceptual experiment using a stimulus set based on real tracks showed a bias towards mirror symmetry over identity. A large fraction of paired random trajectories of few independent steps and turning angles can be expected to exhibit true turn symmetry on the basis of chance alone. But as the number of turning angles increases the number of rigorously symmetrical pairs decreases sharply. Subjects classified nearly 50% of 10-step computer generated random persistent trajectory pairs as symmetrical while only a fraction of a per cent are rigorously so. In fact, they were classified as though they contained only three to four random independent steps. The bias of the human observer and chance may thus account for the perception of a high occurrence of sister trajectory symmetry in vitro.Since branching patterns of sister trajectories do not possess rigorous symmetry, we feel they cannot be taken as evidence of preprogramming. However, the dissimilarity of the structure of sister-cell walks does not rule out the possibility that sister cells are alike in other respects. We found a significant correlation between the time-average velocities of newly born sister cells, and suggest that it and other similarities may be due to the equipartitive character of mitosis. The correlation decays over a 2 · 5-day period after mitosis. But phenotypic sister-cell similarities do not appear to determine the sequence of turning angles in cell walks. If they did a claim might be made for the preprogramming of locomotion.