Pulsatile contractions promote apoptotic cell extrusion in epithelial tissues

Abstract

Extrusion is a mechanism used to eliminate unfit, excess, or dying cells from epithelial tissues. The initial events guiding which cells will be selectively extruded from the epithelium are not well understood. Here, we induced damage in a subset of epithelial cells in the developing zebrafish and used time-lapse imaging to examine cell and cytoskeletal dynamics leading to extrusion. We show that cell extrusion is preceded by actomyosin contractions that are pulsatile. Our data show that pulsatile contractions are induced by a junctional to medial re-localization of myosin. Analysis of cell area during contractions revealed that cells pulsing with the longest duration and highest amplitude undergo progressive area loss and extrude. Although pulses were driven by local increases in tension, damage to many cells promoted an overall decrease in the tensile state of the epithelium. We demonstrate that caspase activation leads to sphingosine-1-phosphate enrichment that controls both tissue tension and pulses to dictate areas of extrusion. These data suggest that the kinetics of pulsatile contractions define a key behavioral difference between extruding and non-extruding cells and are predictive of extrusion. Altogether, our study provides mechanistic insight into how localized changes in physical forces are coordinated to remove defective cells for homeostatic maintenance of living epithelial tissues.

Keywords: cell extrusion; epithelia; physical forces; tissue homeostasis; zebrafish.

Figure 1 –. Pulsatile contractions appear following cellular damage and predict extrusion events

A) Left: Maximum intensity projection (MIP) of a larva tail 2h post-MTZ treatment. Epithelial cells express NTR-mCherry (magenta) and LifeAct-GFP (green). Scale bar = 50µm. Magnified regions are represented by white square. Right: Time-lapse sequence of pulsing epithelial cell until extrusion. White arrowheads indicate pulsing events and red arrowheads extrusion event. Time is in hours and minutes HH:MM. See also Video S1. B) Scatter dot graph correlating the area fold change of pulsing epithelial cells overtime. Dots represent the average area fold change +/− SD of 155 pulses of 29 cells over 9 larvae. Average values of frequency (0.0046s⁻¹), duration (120s), relative area loss (2.5%) and relative amplitude (15%) are represented. See also Figure S1E. C) Scatter dot graph representing the cumulative number of pulses in all epithelial cells (purple line) and the cumulative number of extrusion events (blue line). T = 0h is when live-imaging starts. Analysis was performed in n=5 larvae and dots represent average values +/− SEM. D) Scatter dot graph representing single cell examples of the area fluctuation of an extruding cell, neighboring cell and non-extruding cell. See also Figure S1J. E) Quantification of the duration, relative amplitude, relative area loss and frequency (from left to right) of pulses in extruding, neighboring and non-extruding cells. Analyses were performed on n= 20 cells/category over n= 5 larvae. F) Quantification of the number of extruding cells in larvae treated with 10mM of MTZ for 2h, 10mM of MTZ for 1h or 5mM of MTZ for 2h. Results are expressed as histograms +/− SEM with all values. G) Quantification of the duration, relative amplitude and relative area loss (from left to right) of extruding cells in larvae treated with 10mM of MTZ for 2h (same larvae than in Figure 1E), 10mM of MTZ for 1h and 5mM of MTZ for 2h. Analyses were performed on n= 13–20 cells/category over n= 5 larvae.

Figure 2 –. Pulsatile contractions are a cell-level behavior but induce area changes across neighbors

A) Left: Time-lapse sequence of pulsing epithelial cells expressing LifeAct-GFP. Time is in minutes and seconds MM:SS. Scale bar = 20µm. Third panel represents flow fields where the length of the arrows is proportional to convergence/divergence. Lower panel represents convergence/divergence maps where convergence is blue and divergence is red. Color code gives the amplitude of convergence/divergence in s⁻¹. Dashed red line represents the outline of tracked pulsing cell. Right: Scatter dot graph correlating the area fold change (left axis; grey line as given by segmentation in Tissue Analyzer), divergence from a pulsing cell (right axis; red line) and neighboring cells (right axis; blue line). Dots represent average values of convergence/divergence of 182 pulses +/− SEM in n=3 larvae. B) Time-lapse sequence of epithelial cells expressing LifeAct-GFP. Overlaid color map represents convergence/divergence field where blue is convergence and red is divergence. Red arrow indicates the first pulsing cell and white arrow indicates neighboring cell that pulses afterwards. Scale bar = 20µm. Time is in minutes and seconds MM:SS. See also Video S2 and Figure S2B, S2C. C) Quantification of the duration, relative amplitude and relative area loss of single-cell pulses vs multi-cell pulses. Analyses were performed on n= 10 cells/category over n= 3 larvae. D) Left: Time-lapse sequence of MIPs of larva tails where epithelial cells express NTR-mCherry (magenta) and LifeAct-GFP (green). Dotted white lines represent the contour of the tails. Time is in hours. Scale bar = 100µm. Right: Scatter dot graph correlating the area fold change of DMSO-treated larvae and MTZ-treated larvae. Analysis was performed on n=10 larvae/condition and dots represent average values +/− SEM.

Figure 3 –. Tension at the tissue scale is reduced following damage but increases locally to produce pulsing

A) Quantification of the relative amplitude of pulses in larvae treated with MTZ vs MTZ+Y-27632 (left) or MTZ vs MTZ+cytochalasin (right). See also Video S3 and Figure S3A for analysis of the number of extruding cells. B) Left: Time-lapse sequence of a pulsing cell expressing Myosin II-GFP. Purple and yellow colors represent low and high fluorescence intensities respectively. Time is in minutes and seconds MM:SS. Right: Scatter dot graph correlating the area fold change (left axis; grey line) and myosin fluorescence fluctuation (left axis, red and pink lines) of pulsing cells. Analysis was performed on n=3 larvae and dots represent average values of 79 pulses +/− SEM. Blue arrowheads indicate the peak value of each curve. Green brackets represent the increase in junctional myosin fluorescence at the end of pulsing. See figure S3D for data from single cells. C) Left: Time-lapse sequence of a pulsing cell expressing LifeAct-GFP. Purple and yellow colors represent low and high fluorescence intensities respectively. Time is in minutes and seconds MM:SS. Right: Scatter dot graph correlating the area fold change (left axis; grey line) and F-actin fluorescence fluctuation (right axis; red line) of pulsing cells. Analysis was performed on n=3 larvae and dots represent average values of 182 pulses +/− SEM. Green brackets represent the increase in junctional F-actin fluorescence at the end of pulsing. See figure S3F for data from single cells. D) Time-lapse sequence of a cell expressing LifeAct-GFP where microridges dissipate overtime. Time is in hours and minutes HH:MM. E) Left: MIP of epithelial cells expressing LifeAct-GFP before (green) and after (magenta) laser ablation. Scale bar = 40µm. Red dot represents ablated region. Right: Quantification of the recoil rate in pulsing and non-pulsing regions. F) Left: MIP of epithelial cells expressing LifeAct-GFP before (green) and after (magenta) laser ablation. White and red lines contour cell edges before and after ablation respectively. Scale bar = 40µm. Red dot represents ablated region. Right: Quantification of the recoil rate in DMSO and MTZ-treated larvae. G) Left: 2D slice of epithelial cells expressing NTR-mCherry (magenta). Phosphorylated myosin (P-MLC) is immunostained (green). Scale bar = 20µm. Right: Quantification of junctional P-MLC fluorescence of DMSO and MTZ-treated larvae. Analysis was performed on 10–20 cells/larva averaged over n=5–7 larvae/condition over N=3 separate experiments. H) Quantification of the apparent stiffness of epithelia in DMSO and MTZ-treated larvae. See Figure S3I for representative AFM image.

Figure 4 –. Pulsatile contractions are driven by caspase activity and S1P phosphorylation

A) Upper: MIP of the tail of larvae treated with MTZ overtime. Epithelial cells express NTR-mCherry (magenta) and S1P is immunostained (green). Scale bar = 100µ m. Lower: Scatter dot graph correlating S1P fluorescence of DMSO and MTZ-treated larvae where t = −2h represents addition of MTZ and t = 0 corresponds to the start of microscope acquisition. Analysis was performed on n=30–40 larvae/condition and dots represent average values +/− SEM. See Figure S4A for control analyses of S1P fluorescence. B) Quantification of the relative amplitude of pulses in larvae treated with MTZ vs MTZ+SKI. See also Video S4. C) Quantification of the number of extruding cells in larvae treated with MTZ vs MTZ+SKI. See also Video S4. D) Quantification of the relative amplitude of pulses in NTR vs NTR+Bcl2 overexpressing larvae treated with MTZ. E) Quantification of the number of extruding cells in in NTR vs NTR+Bcl2 overexpressing larvae treated with MTZ. See Figure S4B, S4C for control analyses of S1P fluorescence and caspase 3 positive cells. F) Quantification of the relative amplitude of pulses in larvae treated with MTZ, MTZ+Apoptosis Inhibitor (NS3694), and MTZ+zDVED. See also Video S4. G) Quantification of the number of extruding cells in larvae treated with MTZ, MTZ+Apoptosis Inhibitor (NS3694) and MTZ+zDEVD. See also Video S4. See Figure S4D, S4E for control analyses of S1P fluorescence and caspase 3 positive cells. H) Left: Time-lapse sequence of epithelial cells from larva treated with MTZ and soaked in NucView 405. Epithelial cells express LifeAct-GFP (green) and cleaved caspase 3 is fluorescently labelled (Magenta). Scale bar = 20µm. Time is in hours and minutes HH:MM. See also Video S5. Right: Scatter dot graph representing the number of caspase 3+ cells overtime. Analysis was performed on n=6 larvae and dots represent average values +/− SEM. See Figure S4F for immunofluorescence-based quantification of caspase 3 positive cells.

Figure 5 -. Caspase activation and S1P phosphorylation alter the mechanical state of the tissue following damage

A) Quantification of the apparent stiffness of epithelia of larvae treated with MTZ vs MTZ+SKI. B) Quantification of the apparent stiffness of epithelia in NTR vs NTR+Bcl2 overexpressing larvae treated with MTZ. C) Quantification of the recoil rate in MTZ vs MTZ+SKI treated larvae. D) Quantification of the recoil rate in NTR vs NTR+Bcl2 overexpressing larvae treated with MTZ. E) Quantification of the recoil rate in MTZ, MTZ+Y-27632, or MTZ+cytochalasin treated larvae. F) Quantification of the recoil rate in MTZ vs MTZ+BAPTA treated larvae. G) Quantification of the relative amplitude of pulses in MTZ vs MTZ+BAPTA treated larvae. See also Video S6. H) Quantification of the number of extruding cells in MTZ vs MTZ+BAPTA treated larvae. See also Video S6. I) Upon MTZ treatment, damaged epithelial cells get enriched in S1P. Damage and S1P enrichment lead to decrease in tension within the tissue and induce pulsatile contractions through medio-apical myosin coalescence. Cells that pulse with higher dynamics and are under the highest amount of tension preferentially extrude.