Interestingly global EGFR depletion increased
Interestingly, global EGFR depletion increased the rate of cell elimination everywhere in the notum (Figures 2B–2D), irrespective of the deformation status of the cells. Accordingly, we found that Neurotensin are not any more sensitive to stretching upon EGFR depletion (Figures S6A–S6C; Video S5). This suggests that a ubiquitous basal activity of EGFR is required for cell survival everywhere in the notum, which could be then modulated by tissue deformation. We found that tissue stress and/or compaction can modulate ERK activity and that part of the ERK dynamics correlated with tissue deformations. However, it is very likely that the complex spatiotemporal pattern of ERK activity in the pupal notum (e.g., global downregulation 16 or 17 hr APF and global upregulation 20 hr APF; Video S2) is also controlled by currently unknown patterning genes. EGFR/ERK modulation by deformation may be required to fine tune its activity, to coordinate in time and space cell elimination, and to regulate the number of cells that will be eliminated. A high rate of cell elimination could lead to higher cell spacing and/or an increase of tissue tension, which would feedback negatively on cell elimination through ERK activation. Moreover, the mutual regulation of ERK and mechanics [30, 31, 32] could generate complex temporal dynamics and self-organized properties .
Although we do not know which molecular effectors of ERK pathway are sensitive to mechanical stress, epistasis experiments suggest that modulation occurs upstream and/or at the level of EGFR (see Figures S6A–S6C; Video S5). Accordingly, a large pool of EGFR is located at adherens junctions (Figure S6D) compatible with a modulation of EGFR activity by apical cell geometry and/or mechanical stress. So far, we did not observe obvious variations of the location and concentration of EGFR in the midline compared to the rest of the notum (Figure S6D), arguing for a mechanism based on a modulation of EGFR activity rather than a direct modulation of its concentration/localization. The correlation between tissue stretching or compaction and ERK activity could be explained by different parameters, including membrane tension, cell apical surface, and/or changes in cell volume. The transient ERK downregulation observed after tissue severing and the correlation between compaction rate and ERK activity suggest that ERK is sensitive to strain rate rather than absolute cell size and/or tissue density. Similarly, compressive forces rather than absolute tissue density are responsible for spontaneous MDCK cell elimination . Further exploration of the single cell parameters correlating with ERK fluctuations will help to identify the relevant factors modulating ERK. Although our data support a central role for ERK in modulating cell survival in the pupal notum, we cannot predict fully accurately which cell will engage in apoptosis based on miniCic signal. For instance, we do not know whether caspase activation is triggered by the temporal dynamics or ERK or/and by its absolute levels. Moreover, the probability of cell elimination is still higher in the midline compared to the rest of the tissue upon EGFR depletion (Figures 2C and 2D, 0.75 and 0.42, respectively, versus 0.29 and 0.03 for control clones; Figures 1C and 1D), suggesting that other currently unknown factors also modulate the susceptibility to cell death in the midline. This is in agreement with the absence of ERK downregulation preceding caspase activation in 25% of the midline dying cells (Figure 4). Further work will be required to identify all the factors modulating cell survival in the notum. Finally, the contribution of compaction-induced ERK inhibition to cell elimination and Ras clone expansion may be relevant for other competition scenarios and in pathological conditions. Yki activation in clones was also shown to trigger WT cell deformation and their death in the pupal notum . Cells mutant for the apico-basal polarity protein Scribble are also eliminated through the downregulation of EGFR/ERK , although this is driven in that case by the ligand Sas and the tyrosine phosphatase PTP10D. Finally, compaction-driven ERK downregulation could be relevant for the elimination of misspecified cells in Drosophila wing imaginal disc, which has been associated with an increase of contractility at the clone boundary, leading to cell compaction within the clone . Constitutively active mutant forms of Ras are present in one third of human cancers . Our study suggests that blocking mechanical-induced cell elimination by ERK activation can significantly slow down the expansion of tumoral cells. This opposite role of ERK in tumors (promotion of tumoral cell growth and survival) and the surrounding cells (resistance to mechanical stress) may explain the limited success of Ras/Raf/ERK-targeted cancer therapies .