Non-junctional role of Cadherin3 in cell migration and contact inhibition of locomotion via domain-dependent, opposing regulation of Rac1

Abstract

Classical cadherins are well-known adhesion molecules responsible for physically connecting neighboring cells and signaling this cell-cell contact. Recent studies have suggested novel signaling roles for "non-junctional" cadherins (NJCads); however, the function of cadherin signaling independent of cell-cell contacts remains unknown. In this study, mesendodermal cells and tissues from gastrula stage Xenopus laevis embryos demonstrate that deletion of extracellular domains of Cadherin3 (Cdh3; formerly C-cadherin in Xenopus) disrupts contact inhibition of locomotion. In both bulk Rac1 activity assays and spatio-temporal FRET image analysis, the extracellular and cytoplasmic Cdh3 domains disrupt NJCad signaling and regulate Rac1 activity in opposing directions. Stabilization of the cytoskeleton counteracted this regulation in single cell migration assays. Our study provides novel insights into adhesion-independent signaling by Cadherin3 and its role in regulating single and collective cell migration.

Figure 1

Cell migration assays using Xenopus gastrula stage embryonic mesendoderm; contact inhibition of locomotion (CIL) in collective migration, CIL in single migratory cells, and directionality of single motile cells. (A) Schematic of intravital imaging of mesendoderm closure in Xenopus embryo from stage 11.5. Animal cap ectoderm was removed, and the lip and outer surface of the mesendodermal mantle was placed in contact with a fibronectin-coated cover glass. The right-side image shows the mesendoderm mantle observed with a stereomicroscope. (B) Frames from a confocal time-lapse showing closing mesendoderm mantle expressing membrane-targeted GFP (dotted lines indicate the boundaries). A difference in the expression level of GFP indicates the different origins of opposing sides. (C) Progressive rates of closure from five embryos. Progress of the leading edge from the start time point to closure at each time point is shown. The arrow indicates the time of the collision. After the collision, cell migration stops. (D) Frames from a representative sequence showing lamellipodia retraction. The leading-edge on the darker cell is indicated by yellow arrowheads; the leading edge lamellipodia retracts after touching the brighter opposing cell. (E) Representative frames from a brightfield time-lapse sequence of colliding single mesendodermal cells. The trajectories are shown with blue lines. WT mesendodermal cells change migration direction after the collision. (F) The geometry of single-cell CIL kinetic analysis with a migratory cell (light brown) and an opposing cell (dark brown). (G) Vector plot of a set of wild type (WT) cell collisions (blue arrow—incoming cell; red arrow—mean angle post-collision; dotted circle radius 1—same velocity before and after collision). The mean value of the angle taken by the departing cell (incident angle = 0°) is calculated for a number of collisions (N). Statistical significance is calculated for a circular distribution of post-collision angles (p). (H) Schematic for measurement of directionality (left) and tracked paths of single migrating cells (right). All scale bars are 20 µm. The illustrations were drawn using Adobe Illustrator Version 24.1.1 (https://www.adobe.com/products/illustrator.html).

Figure 2

Truncation mutants of Cdh3 independently regulate collective CIL, single cell CIL, and single cell persistence. (A) Frames from a representative confocal time-lapse sequence (left; leading edge, yellow arrowheads) of membrane-targeted GFP expressing mesendoderm mantle closure in embryos expressing extracellular truncated Cdh3 (ΔE-cdh3), cytoplasmic domain truncated Cdh3 (ΔC-cdh3), and ΔE-cdh3 + ΔC-cdh3. Positions of leading edge mesendoderm movements during closure (including sequences shown at left, n = 5; arrows indicate time of collision). (B) Collisions of ΔE-cdh3, ΔC-cdh3, and ΔE-cdh3 + ΔC-cdh3 expressing single cells (left; trajectories shown in blue). Angle followed by cells after collision (right; mean angle, red; ΔE-cdh3, N = 37, mean = 7.6°, p < 0.001; ΔC-cdh3, N = 40, mean = − 179.8°, p < 0.01; ΔE-cdh3 + ΔC-cdh3, N = 47, mean = − 113.5°, p = 0.07). (C) Collisions between ΔE-cdh3 expressing cells and WT cells. Left: Summary of collisions of ΔE-cdh3 expressing cells into WT cells (N = 32, mean = − 6.7, p < 0.001). Right: Summary of single cell collisions of WT cells into ΔE-cdh3 expressing cells (N = 29, mean = 143.2, p < 0.001). (D) Single cell trajectories over 1 h without collision of ΔE-cdh3, ΔC-cdh3 and ΔE-cdh3 + ΔC-cdh3 expressing cells. (E) Directionality of single cell migration of ΔE-cdh3 (ΔE), ΔC-cdh3 (ΔC), ΔE-cdh3 + ΔC-cdh3 (ΔE + ΔC), ΔE-cdh3 + full length of Cdh3 (FL-cdh3) (ΔE + FL), ΔC-cdh3 + FL-cdh3 (ΔC + FL). All scale bars are 20 µm.

Figure 3

Truncated cadherins modulate Rac1 activity responsible for single cell CIL, collective CIL, and single cell persistence. (A) Rac1 activity is modulated from WT levels by ΔE-cdh3 (ΔE), ΔC-cdh3 (ΔC), and ΔE-cdh3 + ΔC-cdh3 (ΔE + ΔC). (B) Rac1 activity of (A) normalized to WT levels. Rac1 activity in ΔE-cdh3 and ΔC-cdh3 expressing cells was 0.4- and 1.8-fold of WT levels. ΔE + ΔC activity was not significantly different from WT levels. Full-length blots are presented in Figure S7A. (C) Mesendoderm closure in embryos co-injected with ΔE-cdh3 + constitutively active form of Rac1 (caRac). (D) Time-courses of mesendoderm closure from five embryos. (E) Single cell collisions of ΔE-cdh3 + caRac expressing cells. (F) Summary of collisions of ΔE-cdh3 + caRac expressing cells (N = 42, mean = − 167.8°, p = 0.43). (G) Directionality of single cell migrations of ΔE-cdh3 + caRac (ΔE + caRac), ΔC-cdh3 + dominant negative form of Rac1 (dnRac) (ΔC + dnRac), caRac and dnRac + Rac1 morpholino oligomers (RacMO) (dnRac + RacMO) compared with ΔE-cdh3 (ΔE), ΔC-cdh3 (ΔC) and WT, respectively (light red for comparison from Fig. 2E). (H) Single cell collisions of dnRac + RacMO co-injected cells. (I) Summary of collisions of dnRac + RacMO co-injected cells (N = 44, mean = 9.0°, p = 0.08). All scale bars are 20 µm.

Figure 4

Truncated cadherins alter the spatial patterns of Rac1 activity during cell–cell collisions. (A) Schematic of the method for segmenting the cell membrane of migrating cells. The line passing through the center of the cell detects peaks on both sides in the YFP channel and rotates 360° by 1°. (B) Representative Rac1 biosensor FRET image (CFP channel, left) and segmented points for activity quantification (blue line, right). (C) The intensity profile of Rac1 activity along the white dotted line (B; cell boundary indicated). (D) Representative FRET images before, during, and after a typical cell–cell collision of single WT cell. Rac1 activity is shown in pseudocolor; white arrows indicate the directions of cell migrations. (E) Kymograph of Rac1 activity along the cell perimeter during a cell–cell collision with cell initially moving to the top of the frame in (D). The center of the x-axis indicates the front of the migrating cell prior to the collision. The arrowhead and dotted line indicate the time of the collision. After the collision, (F) FRET profile along cell perimeters of five typical cells. The center of the x-axis indicates the front of the migrating cell. (G) Time-course images before, during, and after a typical cell–cell collision of single ΔE-cdh3 expressing cell. (H) Kymograph during the cell–cell collision of ΔE-cdh3 expressing cell. (I) FRET ratio profile along cell perimeters of five typical cells. (J) Time-course images before, during, and after a typical cell–cell collision of single ΔC-cdh3 expressing cell. (K) Kymograph during the cell–cell collision of ΔC-cdh3 expressing cell. (L) FRET ratio profile along cell perimeters of five typical cells. (M) The frequency of WT CIL is high when the first contact occurs in a high Rac1 activity zone and low when contact occurs in a low Rac1 activity zone. (N = 77 contacting at high Rac1 zone, 9 contacting at low Rac1 zone). (N) Percentage of cell perimeter above ratio 1.2 along cell boundary in WT, ΔE-cdh3, and ΔC-cdh3 cells. WT: 12.6 ± 0.18 (mean ± SE)% (N = 57), ΔE-cdh3: 7.6 ± 0.22% (N = 50, p < 0.01), ΔC-cdh3: 27.3 ± 0.26% (N = 59, p < 0.001).

Figure 5

Effects of non-junctional cadherin3 on single-cell CIL and persistence of single-cell migration mitigated by changes in cytoskeletal stability. (A) Summary of collisions of ΔE-cdh3 expressing cells in 20 nM paclitaxel (N = 49, mean = -25, p = 0.71). (B) Summary of collisions of ΔE-cdh3 expressing cells in 20 nM jasplakinolide (N = 37, mean = − 22.9°, p = 0.09). (C) Summary of collisions of WT cells in 20 nM nocodazole (N = 38, mean = 30.7°, p = 0.19). (D) Summary of collisions of WT cells in 20 nM cytochalasin D (N = 42, mean = 7.1°, p = 0.87). (E) The quantified directionality of single-cell migration of ΔE-cdh3 cell in 20 nM Paclitaxel (ΔE in Pac), ΔE-cdh3 cell in 20 nM Jasplakinolide (ΔE in Jas), WT cell in 20 nM Nocodazole (WT in Noc), and WT cell in 20 nM Cytochalasin D (WT in CytD) compared with ΔE-cdh3, and WT, respectively (light red for comparison from Fig. 2E). (F) Model of the hypothetical mechanism regulating CIL and directionality of single-cell migration in cells expressing non-junctional Cdh3 or truncated cdh3s through localized Rac1 activity and cytoskeleton stability. Rac1 is regulated independently by extracellular and cytoplasmic domains of non-junctional cadherin oppositely through unknown factors. Signals induced by cell–cell contact may also regulate CIL with potential signaling through either cis- and trans-bonds between Cdh3. In WT cells, Rac1 activity is held at moderate levels by the extracellular and cytoplasmic domains of non-junctional cdh3. Moderate Rac1 activity leads to moderate levels and gradient of cytoskeletal stability. As a stable cytoskeleton supports persistent migration, the moderate stability of the cytoskeleton in WT cells results in the moderate frequency of change in the direction of cell migration and enables CIL and moderate directionality of single-cell migration. In ΔE-cdh3 expressing cell, expression of the cytoplasmic domain of Cdh3 suppresses Rac1 activity and destabilizing the cytoskeleton; resulting in a low frequency of change in the direction of cell migration, defective-CIL, and lowered persistence. Meanwhile, ΔC-cdh3 hyper activates Rac1 and leads to the stable cytoskeleton and high frequency of change in the direction of cell migration, resulting in frequent turning and cells capable of CIL. The illustration was drawn using Adobe Illustrator Version 24.1.1 (https://www.adobe.com/products/illustrator.html).