Using light and electron microscopy, we have confirmed an earlier observation that chick embryonic pigmented retina epithelial cells (PRE cells) seeded in vitro on cultured sheets of choroid fibroblasts, are able to spread. Spreading is as rapid (and shows the same dependence on lateral contact between PRE cells) as on a serum-coated culture substrate. After 1 h most cells are spreading on the upper surface of the choroid sheet, but after 4 h, some PRE cells can be found sandwiched between overlapping choroid cells, and thus have invaded the sheet. Choroid fibroblasts underlie PRE in vivo, but the ability of PRE cells to spread on cultured fibroblasts is not specific for choroid, since PRE cells spread also on BKH21 hamster kidney fibroblasts, and on fibroblasts from chick embryonic heart. As reported by others for various fibroblastic cells, choroid cells seeded on to choroid sheets or on to cultured PRE are unable to spread. A possible explanation is that spreading of adherent cells is contact-inhibited by the cells in the sheet, just as their leading edges are paralysed on contact, and thus locomotion is inhibited, when fibroblasts collide on a plane substratum. If spreading of seeded cells and cell locomotion are inhibited by the same mechanism, PRE cells should contact-inhibit choroid fibroblasts with which they collide, but not themselves be so inhibited. Using time-lapse cinemicrography, we have found this to be the case. We first established that in homotypic collisions, choroid fibroblasts do show contact inhibition of locomotion, despite the criss-cross (not well monolayered) appearance of confluent cultures. In heterotypic collisions between choroid and PRE we found the predicted nonreciprocal behaviour: the choroid leading edge is paralysed on collision, and the cell subsequently retracts, whereas the active PRE margin appears to be completely unaffected. Speed measurements from a series of such collisions show that the speed of choroid cells is markedly reduced on collision with PRE, whereas the slight slowing of PRE is not statistically significant. We have observed similar behaviour in heterotypic collisions between various epithelial and fibroblastic cells, and so it seems possible that non-reciprocity may prove general for this interaction. If so, it has important implications for the role of contact inhibition of locomotion in phenomena such as morphogenesis, wound healing and the invasiveness of carcinoma cells. On the one hand, non-reciprocal contact-inhibition of locomotion may permit the spreading of epithelia over mesenchymal cells, thus generating or restoring an epithelial bounding membrane. On the other hand, in the absence of other interactions, it would fail to inhibit the invasion of mesenchymal territory by aberrant epithelial cells, or presumably by epithelial free edges.

This content is only available via PDF.