In most cancers, transformation begins in a single cell in an epithelial cell sheet. However, it is not known what happens at the interface between non-transformed (normal) and transformed cells once the initial transformation has occurred. Using Madin-Darby canine kidney (MDCK) epithelial cells that express constitutively active, oncogenic Ras (Ras(V12)) in a tetracycline-inducible system, we investigated the cellular processes arising at the interface between normal and transformed cells. We show that two independent phenomena occur in a non-cell-autonomous manner: when surrounded by normal cells, Ras(V12) cells are either apically extruded from the monolayer, or form dynamic basal protrusions and invade the basal matrix. Neither apical extrusion nor basal protrusion formation is observed when Ras(V12) cells are surrounded by other Ras(V12) cells. We show that Cdc42 and ROCK (also known as Rho kinase) have vital roles in these processes. We also demonstrate that E-cadherin knockdown in normal cells surrounding Ras(V12) cells reduces the frequency of apical extrusion, while promoting basal protrusion formation and invasion. These results indicate that Ras(V12)-transformed cells are able to recognize differences between normal and transformed cells, and consequently leave epithelial sheets either apically or basally, in a cell-context-dependent manner.

In the phagocytosis field, ROS production by the phagocyte NOX has been associated with pathogen killing for the last 50 years. Since the discovery of nonphagocyte NOX, numerous other roles for ROS production have been identified. Oxidative stress and ROS-mediated signaling have received much attention in recent years. Much lower concentrations of ROS may be required for signaling compared with microbial killing. Based on the discoveries in nonphagocytic cells, it became logical to look for ROS functions distinct from pathogen killing, even in phagocytes. ROS are now linked to various forms of cell death, to chemotaxis, and to numerous modifications of cellular processes, including the NOX itself. ROS functions are clearly concentration-dependent over a wide range of concentrations. How much is required for which function? Which species are required for how much time? Is ROS signaling only a side effect of bactericidal ROS production? One major obstacle to answer these questions is the difficulty of reliable quantitative ROS detection. Signal transduction often takes place on a subcellular scale over periods of seconds or minutes, so the detection methods need to provide appropriate time and space resolution. We present examples of local ROS production, decreased degradation, signaling events, and potentially ROS-sensitive functions. We attempt to illustrate the current limitations for quantitative spatiotemporal ROS detection and point out directions for ongoing development. Probes for localized ROS detection and for combined detection of ROS, together with protein localization or other cellular parameters, are constantly improved.