Invasive locomotory behaviour between malignant human melanoma cells and normal fibroblasts filmed in vitro.

Explants of human malignant melanoma (MM96), normal adult human skin fibroblasts (HSF) and embryonic chick ventricle were confronted in pairs. The 2 outwandering populations in each confrontation eventually met in a situation where each could potentially invade the other. The human explants were artificially prepared from dissociated cells. The primary objective of the study was to compare the relative invasive capacities of the malignant and nonmalignant human cell populations against a standard population of chick heart fibroblasts (CHF). Relative invasiveness was also compared for (a) malignant human melanoma cells against human and avian fibroblasts; (b) the 2 different fibroblast populations against MM96; and (c) the 2 fibroblast populations against each other. Time-lapse films were prepared for each confrontation. Each cell population was also filmed in a free or unconfronted state. The films were analysed in terms of (a) cell speeds in relation to numbers of contacts; (b) the net radial outward velocity; (c) directional frequencies of movements; and (d) cell behavior resulting from heterologous collisions. Replicate cultures were fixed and stained 24--40 h after junction. Measurements from these indicated relative distances travelled by each population towards the opposing population and towards free space. Nuclear overlap and population density measurements were also recorded. MM96 cells invaded the standard chick fibroblast population but eventually encountered moderate obstruction. Invasion was mainly by oriented movement between successive pairs of chick cells. Contact paralysis of ruffling of MM96 was not seen. Human fibroblasts did not invade the CHF population. They were contact inhibited by chick cells and typically reversed their direction of movement. MM96 cells invaded the human fibroblast population without obstruction. They were not usually inhibited by HSF and several different pathways for locomotory invasion, including overlapping, were involved. Cells of both fibroblast populations were contact inhibited by leading-edge collisions with MM96 cells. Fibroblast invasion of MM96 was by underlapping of orthogonally oriented, bipolar melanocytes or dendritic processes. Human fibroblasts were less obstructed than chick fibroblasts during invasion of MM96. Initiation of invasion of HSF by chick cells was apparently orientation dependent. Invasion by CHF was assisted by the alignment, morphology and contact-mediated withdrawal of the human cells. Heterologous contact inhibition was not apparently defective for either cell type. In all confrontations involving both malignant and non-malignant cells, the extent and pattern of invasion could be related to apparent deficiencies in heterologous contact inhibition of locomotion.