Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 161 (2), 361-73

Inter-cellular Forces Orchestrate Contact Inhibition of Locomotion

Affiliations

Inter-cellular Forces Orchestrate Contact Inhibition of Locomotion

John R Davis et al. Cell.

Abstract

Contact inhibition of locomotion (CIL) is a multifaceted process that causes many cell types to repel each other upon collision. During development, this seemingly uncoordinated reaction is a critical driver of cellular dispersion within embryonic tissues. Here, we show that Drosophila hemocytes require a precisely orchestrated CIL response for their developmental dispersal. Hemocyte collision and subsequent repulsion involves a stereotyped sequence of kinematic stages that are modulated by global changes in cytoskeletal dynamics. Tracking actin retrograde flow within hemocytes in vivo reveals synchronous reorganization of colliding actin networks through engagement of an inter-cellular adhesion. This inter-cellular actin-clutch leads to a subsequent build-up in lamellar tension, triggering the development of a transient stress fiber, which orchestrates cellular repulsion. Our findings reveal that the physical coupling of the flowing actin networks during CIL acts as a mechanotransducer, allowing cells to haptically sense each other and coordinate their behaviors.

Figures

None
Figure 1
Figure 1
Hemocyte Contact Inhibition Involves Multiple Stages that Are Synchronous and Coordinated in Colliding Partners (A) Dispersal of hemocytes labeled with a nuclear marker (red) beneath the ventral surface of a Drosophila embryo (bright-field) at developmental stages 14 and 15. (B) Automatic tracking of nuclei (red) of colliding hemocytes while also registering collisions with microtubules (green). Time point of microtubule alignment (arrowhead) allows for temporal registration of CIL events in subsequent kinematic analyses. (C) Time course of hemocyte accelerations (black arrows) surrounding a collision event with reference to the colliding partner (red arrow). All time points show random accelerations except at −120, 0, and 180 s where there is a bias along the x axis. p < 0.05, ∗∗∗p < 0.001. (D) Graph showing the internuclear distance of colliding cells during the CIL time course. Note the change in slope at −120, 0, and 120 s. Error bars represent SD. (E) Graph showing nuclear speed during collisions. Note the increase in speed at −120 s and the subsequent decrease upon microtubule alignment. Error bars represent SD. (F) Time-lapse sequence of colliding hemocytes labeled with an F-actin (magenta) and a microtubule (green) probe. Arrows highlight region of lamellae overlap. (G) Kymograph of the region highlighted in (F) showing the time course of actin fiber formation (arrowhead highlights the initial development of the actin fiber) and microtubule alignment. (H) Kymograph of lamellar activity (red regions show lamellar retraction and blue extension) in colliding partners along the actin fiber (red dotted line in schematic). Note that retraction is simultaneous in colliding cells upon lamellae release. (I) Quantification of the rate of lamella retraction over time. Error bars represent SEM. (J) Quantification of lamella retraction rates at 5 and 20 s after cell separation compared with average retraction rates in freely moving cells. Error bars represent SD. p < 0.05. See also Figure S1 and Movie S1.
Figure 2
Figure 2
During Contact Inhibition, the Actin Network Is Rapidly Reorganized in Colliding Partners (A) Top panels are still images from a time-lapse movie of hemocytes containing labeled F-actin during a collision. While cells are in contact, an actin fiber develops between the cell body and the point of contact in colliding partners (arrowheads), which often deforms and breaks upon lamellar retraction (red arrow). Bottom panels highlight actin flow dynamics obtained from the pseudo-speckle analysis. Note that the decreased actin flow in the vicinity of lamellae overlap (highlighted by yellow arrows) is due to the inability of the algorithm to distinguish between the two networks. (B) Kymograph of the region surrounding the actin fiber highlighting the actin retrograde flow dynamics and the alignment of the microtubule bundles (pseudocolored white). (C and D) Instantaneous changes in retrograde flow rate quantified from lamellae contact (C) or lamellae separation (D). (E and F) Instantaneous changes in retrograde flow direction quantified from lamellae contact (E) or lamellae separation (F). For (C)–(F), error bars represent SD. See also Figure S2 and Movie S2.
Figure 3
Figure 3
Actin Network Reorganization Correlates with the Formation of a Transient Cell-Cell Adhesion (A) Still image of a collision between hemocytes expressing mCherry-Zyxin (green) and labeled F-actin (magenta), which highlights the inter-cellular adhesion at the point of initial contact (arrowhead). Arrows highlight region of lamellae overlap. (B) Kymograph of Zyxin and actin dynamics in the region of the actin fiber. Note that the punctum of Zyxin forms in line with the actin fiber and persists for the duration of the time in contact (arrowhead highlights the initial formation of the punctum). (C) Quantification of the maximum intensity of Zyxin and average actin flow rate during the collision. (D) Analysis of actin flow dynamics in comparison with Zyxin localization (pseudocolored white). Note that the region of low retrograde flow develops in line with the inter-cellular adhesion (arrows). (E) Kymograph of Zyxin and microtubule dynamics in the region of the actin fiber highlighting microtubule targeting of the Zyxin puncta. (F) Maximum intensity of Zyxin and microtubules at the inter-cellular adhesion in the region highlighted in (E). (G–J) Cross correlation of the instantaneous changes in flow rate (G and H) and flow direction (I and J) in lamellae of colliding cells. Error bars represent SEM. Red dotted lines represent the mean correlation between colliding cells immediately prior to cell-cell contact with the thickness representing the SEM. See also Movie S3.
Figure 4
Figure 4
Lamellar Stresses Are Increased and Redistributed during CIL (A) Kymographs of lamellar recoil upon laser abscission of the actin network in freely moving and colliding cells. Dotted rectangle highlights the width of the ablation region. (B) Quantification of recoil rate over time and initial recoil rate upon laser abscission. Error bars represent SEM. ∗∗p < 0.01. (C) Quantification of lamellae strain over time upon laser abscission and modeled forces assuming that the actin network behaves elastically over short time scales. The elastic and dissipative mechanical properties in the lamellae are modeled by an exponential decay of the strain that is overlaid onto the constant retrograde flow. Note that zero strain represents the end of the exponential decay. Assuming mechanical properties similar to previously published lamellae we can estimate the tension. Inset: Sketch illustrating the mechanical model of an elastic and dissipative element. The strain u is calculated by the ratio Δl /l. Error bars represent SEM. ∗∗p < 0.01. (D) Hemocyte velocities in freely moving and colliding cells 60 s after laser abscission with respect to the ablation site (red arrow). Magenta arrow is the average direction of the population. Note that after mock ablation there was a significant forward movement of cells, while ablation of the fiber during cell collision led to a significant rearward movement. p < 0.05. (E) Localization of actin network stress during cell collision. Top panels: a time-lapse series of a hemocyte containing labeled F-actin undergoing a collision (adapted from Figure 2A). Bottom panels: modeled intracellular actin stresses. Note that stresses were only measured for regions of the lamella that persisted for a 40-s period as deformation history is required in the analysis. Arrows highlight region of lamellae overlap. Dotted line highlights the redistribution of stresses around the cell body and asterisks the regions of high stress that colocalize with the actin fiber. (F) Kymograph of lamellar stresses over the region that colocalized with the actin fiber. Note the redistribution of stress from the back of the network to the front. (G) Kymograph of the instantaneous changes in actin flow direction in the region colocalizing with the actin fiber. (H) Quantification of the mean change in flow direction of the actin network in three regions corresponding to the back, middle, and front of the actin fiber. Note that the changes initially increase in the rear of the network. See also Movie S4.
Figure 5
Figure 5
The Actin Fiber that Couples Colliding Cells Is a Stress Fiber-like Structure (A) Still image of a freely moving hemocyte containing labeled F-actin (magenta) and Myosin II (green). (B) Quantification of Myosin II tracks in freely moving cells. (C) Still image of a collision between hemocytes containing labeled actin and Myosin II. Note the puncta of Myosin II along the actin fiber (inset). Arrows highlight region of lamellae overlap. (D) Quantification of Myosin II tracks for 40 s upon lamellae overlap during CIL. (E) Kymograph of the region surrounding the actin fiber in (C) highlighting Myosin II accumulation during a collision. (F) Quantification of the increase in actin and Myosin II intensity in the region corresponding to the actin fiber relative to values prior to lamellae contact. Error bars represent SEM. (G) Quantification of Myosin II intensity in regions corresponding to the back versus the front of the actin fiber during CIL. (H) Analysis of actin flow dynamics in comparison with Myosin II localization (pseudocolored white). Note that actin network reorganization precedes Myosin II accumulation along the stress fiber. (I) Still image of a collision between hemocytes containing labeled actin (magenta) and Diaphanous (green). Arrows highlight region of lamellae overlap. (J) Kymograph of the region surrounding the actin fiber in (I). See also Figure S3 and Movie S5.
Figure 6
Figure 6
A Stress Fiber-like Structure Is Required for a Normal CIL Response (A–C) Myosin II mutant (zip1) collisions. (A) Top panels are still images from a time-lapse movie of hemocytes containing labeled F-actin during a collision. Bottom panels are heatmaps obtained from the pseudo-speckle analysis showing no substantial changes in retrograde flow. Arrows highlight region of lamellae overlap. (B) Kymograph of lamellar activity in colliding partners in a region perpendicular to the point of cell contact (red regions highlight lamellar retraction and blue extension). (C) The speed of lamellar retraction in myosin II mutants was quantified at the time of separation to reveal that the retraction rate was no different to freely moving cells. Error bars represent SD. p < 0.05. Note that control retraction rates are from Figure 1J. (D) Quantification of actin fiber formation in control, zip1 and dia5 mutant hemocytes during CIL. The graph represents the relative increase in actin intensity within the region encompassing the actin fiber (red box in schematic) with respect to the surrounding regions of the actin network (blue boxes in schematic). (E) Quantification of the cessation of forward movement during CIL in which the mean distance between the initial point of contact and the nucleus was measured and compared to the distance at the time of cell separation. This analysis revealed that the zip1 and dia5 mutants failed to inhibit their forward motion in comparison to control cells. Error bars represent SD. p < 0.05. (F) Graph of mean time of lamellae contact revealed that zip1 and dia5 mutants maintained cell-cell contacts for a longer duration than control cells. Error bars represent S.D. ∗∗p < 0.01. (G–I) diaphanous mutant (dia5) collisions analyzed as in (A), (B), and (C). See also Figures S4, S5, and S6 and Movie S6.
Figure 7
Figure 7
A Coordinated CIL Response Is Necessary Hemocyte Patterning (A) Time course of hemocyte accelerations in dia5 mutants (black arrows) surrounding a collision event with reference to the colliding partner (red arrow). All time points show random accelerations except the time of microtubule alignment. ∗∗p < 0.01. (B) Quantification of average cell direction during the CIL time course as highlighted in the schematic. Blue highlights forward movement and red movement away from the colliding partner. Error bars represent SD. (C) Cell velocities at 240 s after microtubule alignment with respect to the colliding partner (red arrow). Magenta arrows are the resultant velocities. Note that only controls show a significant movement away from the colliding partner. p < 0.05. (D) The average regions occupied by hemocytes during their developmental dispersal revealed a disruption in the even spacing in diaphanous mutants. (E) Tracks of hemocytes migrating over a 20-min period after they have spread throughout the embryo. (F) Quantification of the maximum distance hemocytes migrate from the tracks measured in (E) revealed that dia5 mutants migrate over greater distances in the embryo. ∗∗∗p < 0.001. See also Figure S7 and Movie S7.
Figure S1
Figure S1
Hemocyte Contact Inhibition Is Precisely Controlled, Related to Figure 1 (A) Time course of hemocyte accelerations at 20 s intervals (black arrows) surrounding a collision event with reference to the colliding partner (red arrow). Note the significant back acceleration only upon microtubule alignment. ∗∗∗p < 0.001. (B) Plot of the SD from the analysis of internuclear distance over time during CIL (Figure 1D). Note that at −120 s the variance in the data decreased suggesting that the process was becoming regulated at this stage. Immediately upon microtubule alignment, the variance suddenly reduced again highlighting that microtubules were tightly controlling the spacing between colliding cells. (C) Example of a hemocyte collision in which regions of lamellae extension (blue) and retraction (red) were highlighted. The lamellae of colliding cells were divided up into regions facing the colliding partner (front) or regions facing away from the colliding partner (back) and subsequently quantified. (D) Average area of lamellar extension and retraction in the regions highlighted in (C). Note that during CIL, lamellar retraction at the front precedes new lamellae formation at the back of the cell. Error bars represent SEM.
Figure S2
Figure S2
Actin Dynamics in Freely Moving and Colliding Hemocytes, Related to Figure 2 (A) A freely moving hemocyte containing labeled F-actin (top panel) analyzed by pseudo-speckle microscopy. Middle panel highlights the actin retrograde flow direction while the bottom panel shows both the magnitude and direction of the flow. (B) Analysis of F-actin intensity within a corridor that corresponds to the region of actin fiber formation during contact inhibition as highlighted in the schematic. (C) Analysis of the retrograde flow rate during CIL highlighting the reduction and subsequent increase in actin flow upon lamellae separation. (D) Kymograph of a region perpendicular to the actin fiber (red dashed line in schematic) highlighting the recruitment of preexisting filamentous actin within the lamella. Note the increased rate of recruitment of actin fibers at the periphery (yellow dashed lines) compared with those closer to the center of the lamella (green dashed lines).
Figure S3
Figure S3
Myosin II and Diaphanous Localization in Hemocytes, Related to Figure 5 (A) Top panel shows a still image of a lamella of a freely moving hemocyte containing labeled F-actin. Bottom panel highlights tracked Myosin II particles within the lamella. The Myosin tracks were color coded in time to highlight their retrograde movement such that the first time point is labeled red, second time point green, and the third blue. (B) Distribution of Myosin II particle speeds in freely moving hemocytes revealed a mean speed of 3.4 ± 0.6 μm/min. (C) Quantification of the lateral displacement of Myosin II particles from the tracks in Figures 5B and 5D revealed an increase in horizontal displacement of Myosin II toward the actin fiber during cell collision. ∗∗∗p < 0.001. (D) Still image of a colliding hemocyte containing labeled F-actin (magenta) and Myosin II (green). Yellow arrows highlight the region of lamellae overlap. The inset shows the region of the actin fiber used for the line scan analysis. Note the repeating peaks of Myosin II intensity along the fiber. (E) Still images of control and constitutively active Diaphanous expressing hemocytes with labeled F-actin (magenta) and Myosin II (green). Note the decrease in area of hemocytes containing constitutively active Diaphanous and the enhanced Myosin II localization within the lamella (arrow). (F) Freely moving hemocyte containing labeled wild-type Diaphanous (green) and F-actin (magenta). (G) Line scan of the region highlighted in (F) revealed a peak of Diaphanous at the ends of filopodia.
Figure S4
Figure S4
Myosin II Drives Retrograde Flow in Hemocytes, Related to Figure 6 (A) Still images of hemocyte dispersal along the ventral surface of Stage 15 embryos. Hemocyte dispersal was disrupted in myosin II mutant (zip1) embryos and expression of GFP-tagged Myosin II specifically in myosin II mutants rescued dispersal. (B) The mean retrograde flow across the lamella of freely moving cells was calculated for control, myosin II mutant (zip1), and myosin II mutant rescue hemocytes. p < 0.05. (C) Probability density function (PDF) of retrograde flow rates in freely moving control, myosin II mutant (zip1), and myosin II mutant rescue hemocytes revealed that expression of GFP-tagged Myosin II increased the distribution of higher flow values in the zip1 mutant cells.
Figure S5
Figure S5
Migration of diaphanous Mutant Hemocytes, Related to Figure 6 (A) Left panel is a still image of a freely moving hemocyte containing labeled F-actin. Right panel is a heatmap of the actin retrograde flow field. (B) The mean retrograde flow across the lamella of freely moving cells was calculated for control, and diaphanous mutant (dia5) hemocytes. (C) Probability density function (PDF) of retrograde flow rates in freely moving control and diaphanous mutant hemocytes revealed a similar distribution. (D) Quantification of cell speed in freely moving control and diaphanous mutant hemocytes. Error bars represent SEM. (E) Quantification of the directional ratio (persistence) of freely moving control and diaphanous mutant hemocytes. Error bars represent SEM. (F) Quantification of instantaneous changes in flow rate in control and diaphanous mutant hemocytes during contact inhibition. Note that the control analysis was taken from Figure 2C. Error bars represent SD. (G) Cross correlation of the instantaneous changes in flow rate in lamellae of colliding diaphanous mutant cells. Error bars represent SEM. Red dotted lines represent the mean correlation between colliding cells immediately prior to cell-cell contact with the thickness representing the SEM. Note that there is no increase in the correlation coefficient upon lamellae contact as observed in control cells (Figure 3G). (H) Quantification of instantaneous changes in flow direction in control and diaphanous mutant hemocytes during contact inhibition. Note that the control analysis was taken from Figure 2E. Error bars represent SD. (I) Cross correlation of the instantaneous changes in flow direction in lamellae of colliding diaphanous mutant cells. Error bars represent SEM. Red dotted lines represent the mean correlation between colliding cells immediately prior to cell-cell contact with the thickness representing the SEM. Note that there is no increase in the correlation coefficient upon lamellae contact as observed in control cells (Figure 3I).
Figure S6
Figure S6
Analysis of Zyxin and Microtubule Dynamics in diaphanous Mutant Collisions, Related to Figure 6 (A) Localization of Zyxin in diaphanous mutant hemocytes during a collision. Note the puncta of Zyxin (arrowhead) at the region of lamellae overlap. (B) Kymograph of Zyxin puncta during a diaphanous mutant collision. Note the presence of Zyxin at the region of lamellae overlap throughout the time course of the response. (C) A collision between diaphanous mutant hemocytes containing labeled F-actin (magenta) and microtubules (green). (D) Kymograph of the region highlighted in (C) showing the alignment between the microtubule networks (arrowhead) in diaphanous mutant cells.
Figure S7
Figure S7
Aberrant CIL in diaphanous Mutant Cells Disrupts Hemocyte Spacing, Related to Figure 7 (A) Time course of dia5 mutant hemocyte accelerations at 20 s intervals (black arrows) surrounding a collision event with reference to the colliding partner (red arrow). Note that unlike controls (Figure S1A) there is no significant back acceleration upon microtubule alignment. (B) Analysis of internuclear distance over time during the CIL response in control (blue line) and diaphanous mutant cells (green line). Note the reduced capacity of diaphanous mutant cells to slow down upon microtubule alignment and subsequently separate. Control analysis was taken from Figure 1D. (C) Still image of wild-type hemocytes dispersed within the ventral surface of a Stage 15 Drosophila embryo containing labeled F-actin (green) and nuclei (magenta). (D) Still image of diaphanous mutant hemocytes dispersed within the ventral surface of a Stage 15 Drosophila embryo containing labeled F-actin (green) and nuclei (magenta). Note the clumping of hemocytes (red arrows) with large regions devoid of cells (dotted line). (E) Still image of a control simulation (as generated in Davis et al. [2012]) showing the even spacing of cells (left panel) which results in an evenly spaced domain map (right panel). (F) Still image of a simulation in which the cells fail to consistently take into account in the direction of the colliding partner during CIL. Note the aberrant spacing of cells (left panel), which results in a failure to acquire a defined domain map (right panel).

Similar articles

See all similar articles

Cited by 32 PubMed Central articles

See all "Cited by" articles

References

    1. Abercrombie M. Control mechanisms in cancer. Eur. J. Cancer. 1970;6:7–13. - PubMed
    1. Abercrombie M., Heaysman J.E. Observations on the social behaviour of cells in tissue culture. I. Speed of movement of chick heart fibroblasts in relation to their mutual contacts. Exp. Cell Res. 1953;5:111–131. - PubMed
    1. Adams C.L., Chen Y.T., Smith S.J., Nelson W.J. Mechanisms of epithelial cell-cell adhesion and cell compaction revealed by high-resolution tracking of E-cadherin-green fluorescent protein. J. Cell Biol. 1998;142:1105–1119. - PMC - PubMed
    1. Astin J.W., Batson J., Kadir S., Charlet J., Persad R.A., Gillatt D., Oxley J.D., Nobes C.D. Competition amongst Eph receptors regulates contact inhibition of locomotion and invasiveness in prostate cancer cells. Nat. Cell Biol. 2010;12:1194–1204. - PubMed
    1. Betz T., Koch D., Lim D., Käs J.A. Stochastic actin polymerization and steady retrograde flow determine growth cone advancement. Biophys. J. 2009;96:5130–5138. - PMC - PubMed

Supplemental References

    1. Franke J.D., Montague R.A., Kiehart D.P. Nonmuscle myosin II generates forces that transmit tension and drive contraction in multiple tissues during dorsal closure. Curr. Biol. 2005;15:2208–2221. - PubMed

Publication types

LinkOut - more resources

Feedback