Tissue geometry and mechanochemical feedback initiate rotational migration in Drosophila

Abstract

Collective migration of epithelial cells drives diverse tissue remodeling processes. In many cases, a free tissue edge polarizes the cells to promote directed motion, but how edge-free or closed epithelia initiate migration remains unclear. Here, we show that the rotational migration of follicular epithelial cells in the Drosophila egg chamber is a self-organizing process. Combining experiments and theoretical modeling, we identify a positive feedback loop in which the mechanosensitive behavior of the atypical cadherin Fat2 synergizes with the rigid-body dynamics of the egg chamber to both initiate and sustain rotation. Mechanical constraints arising from cell-cell interactions and tissue geometry further align this motion around the egg chamber's anterior-posterior axis. Our findings reveal a biophysical mechanism - combining Fat2-mediated velocity-polarity alignment, rigid-body dynamics, and tissue geometry - by which a closed epithelial tissue self-organizes into persistent, large-scale rotational migration in vivo, expanding current flocking theories.

Figure 1:. Egg chamber rotation begins within the germarium.

A) Illustration of a transverse section through a developmental array of egg chambers (ovariole). The boxed region highlights the organization of the follicular epithelium. B) Illustration of rotational epithelial migration (arrow), which is driven by follicle cell crawling along the stationary basement membrane. C) Illustration of the basal epithelial surface showing how Fat2 acts in trans from the trailing edge of each cell to stabilize WRC activity at the leading edge of the cell behind. D) Illustrations of transverse sections through stage 1 and 2 egg chambers, highlighting the morphological differences between them. E) Movie stills of a transverse section through one egg chamber that capture the transition from stage 1 to 2. Pink dashed line shows the location of the line-scans used to generate the plot in G. Nuclei marked with SpyDNA. F) Movie stills focused on the follicular epithelium of the same egg chamber as in E. Two different cells are pseudocolored in each row of images to show cell movement over 30 minutes at each of the three time points. G) Line-scans of Col IV-GFP through the anterior-most pole of the follicular epithelium for 5 different rotating egg chambers at stage 1. The bimodal distribution indicates that the BM has not yet formed over the egg chamber’s anterior. H) Quantification of follicle cell migration rates for egg chambers at stage 1. Bar represents mean ± SD. I) Movie stills of stage 1 egg chambers with corresponding kymographs that show either constant migration or the onset of migration (arrowhead). Scale bars, 10μm.

Figure 2:. Fat2 can mediate symmetry-breaking at multiple developmental stages.

A) Representative images of Fat2-3xGFP at the basal surface of the follicular epithelium. B) Quantification of Fat2-3xGFP intensity at basal surface cell edges. 109-30>fat2-RNAi depletes Fat2 from the follicular epithelium at stages 2 and 3, but Fat2 levels are equivalent to controls by stage 7. Each data point represents one egg chamber. C) Representative centroid tracks of follicle cell movement over 10 minutes. D) Quantification of follicle cell migration rates over developmental time. 109-30>fat2-RNAi blocks migration at early stages but migration is equivalent to controls by stage 7. Each data point represents one egg chamber. In order on graph, $n=6,6,5,10,10,9,11,11,12,16,19,10,11,9,9$. E) Movie stills of stage 5 egg chambers with corresponding kymographs that show either constant migration or delayed onset of migration (arrowhead). F) Illustration summarizing how 109-30>fat2-RNAi delays the onset of migration. For panels B and D, Two-way ANOVA with Tukey’s multiple comparisons test; ns, p > 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Bars represent mean ± SD. Scale bars, 10μm.

Figure 3:. Fat2 becomes planar polarized concurrent with the onset of rotation.

A) Representative images of Fat2-3xGFP at the basal surface of the follicular epithelium. B) Quantification of Fat2 polarization to leading-trailing cell-cell interfaces. Each data point represents the ratio of Fat2-3xGFP brightness at leading-trailing (horizontal) versus side (vertical) cell edges in one epithelium. Fat2 goes from unpolarized to polarized in 109-30>Abi-RNAi with roughly the same timing as the onset of rotation in this background. Two-way ANOVA with Tukey’s multiple comparisons test; ns, p > 0.05, *p < 0.05, **p < 0.01, ****p < 0.0001. Bars represent mean ± SD. C) Illustration of non-migrating cells with uniformly distributed Fat2, versus migrating follicle cells with polarized Fat2. Scale bar, 10μm.

Figure 4:. Fat2 promotes local motility at the basal surface before rotation begins.

A) Representative images of a 109-30>fat2-RNAi epithelium that is rotating, a 109-30>fat2-RNAi epithelium pre-rotation, and a fat2^N103-2 epithelium that will never rotate. Panels on the left show the overlay of the first (green) and last (magenta) frame of a 30-min movie (Movie 6). Panels on the right show cell centroid tracks from the same movie. B) Quantification of cell centroid displacement at apical versus basal surfaces in 109-30>fat2-RNAi epithelia pre-rotation $(n=6)$ or that are already rotating $(n=5)$. Local motility at the basal surface precedes the onset of rotation. Each data point represents the average of cell displacements in one epithelium. Solid lines connect measurements from the same egg chamber. Dotted line at represents the average cell displacements from fat2^N103-2 samples (Fig. S2). Paired t-test; ns, p > 0.05, **p < 0.01. C) Diagram denoting apical and basal imaging planes. D) Representative cell centroid tracks taken at the basal surfaces of five epithelia. The polar order parameter increases from left to right. E) Quantification of the polar order parameter for basal surfaces tracks of 109-30>fat2-RNAi epithelia that are either rotating (red, open circles; $n=6$) or pre-rotation (red, closed circles; $n=24$), compared to fat2^N103-2 epithelia (yellow circles; $n=13$). Some pre-rotation epithelia have polar order parameters that exceed those of rotating epithelia, showing that the follicle cells can align their basal surface movements before rotation begins. Scale bars, 10μm.

Figure 5:. Rigid-body dynamics and the mechanosensitive behavior of Fat2 can initiate rotation.

A) To model the delayed migration assay, the egg chamber is represented as an ellipsoidal rigid body with eccentricity $e$. The dimensions of the egg chamber are rescaled with the radius of the equatorial cross section. B) The crawling force $b_i$ models protrusive activity at the basal epithelial surface and the viscous drag $f_i$ models the interaction between the epithelium and the BM for cell $i$. C) The Fat2 distribution is modeled by $c_i(\beta ,t)$ where $\beta$ is the angular direction in the tangent space. D) Fat2 dynamics are mechanosensitive, whereby Fat2 is recruited to each cell’s trailing edge with respect to the migration direction $v_i$. E) The model breaks rotational symmetry, generating either clockwise ${(\omega }_3<0)$ or counter-clockwise ${(\omega }_3<0)$ rotation about the AP axis. $N=1000$ model trajectories starting from an isotropic initial condition ($c_i(\beta ,0)=1/2\pi ,b_i=0$). F) Representative model trajectory with symmetry broken, resulting in sustained counterclockwise rotations and the corresponding evolution of the cell-averaged Fat2 distribution $⟨c_i(\beta ,t)⟩$. The Fat2 distribution evolves from an isotropic configuration to a polar configuration centered around $\beta =-\pi /2$. The time evolution of the quantities in F is given in Movie 7.

Figure 6:. Mechanical constraints specify the rotational axis.

A) Illustration of a transverse section through a rotating egg chamber at stage 1, highlighting the shape of the interface between a subset of pre-stalk cells (gold) and the anterior-most follicle cells (magenta) in the germarium. B) Movie stills of maximum intensity projections of nuclei generated from a 2.5-hour movie (Movie 8). Stationary pre-stalk cell nuclei are pseudocolored gold, migrating anterior follicle cell nuclei are pseudocolored magenta. The overlay shows the change in the cells positions at these time points. C) The polar angle ${\theta }_s$ quantifies the extent of contact between the pre-stalk cells and the egg chamber. The unit vector $\stackrel{ˆ}{{\mathbf{\text{e}}}_3}$ is along the pre-stalk axis, which aligns with the egg chamber’s AP axis. D) The motion of the egg chamber tangential (normal) to the prestalk/egg chamber interface generates viscous (elastic) resisting forces. Rotational motion of the egg chamber is described in the inertial (i.e. fixed) frame $E:\left\{\stackrel{ˆ}{{\mathbf{\text{e}}}_1},\stackrel{ˆ}{{\mathbf{\text{e}}}_2},\stackrel{ˆ}{{\mathbf{\text{e}}}_3}\right\}$. The orientation of a vector is represented in spherical coordinates $(\theta ,\varphi )$. E) The rotational symmetry-breaking dynamics of a stage 1 egg chamber is described using $\left|\omega \right|,\left|{\omega }_3\right|=\left|\omega \cdot \stackrel{ˆ}{{\mathbf{\text{d}}}_3}\right|$ and average Fat2 distribution $⟨c_i(\beta ,t)⟩$. F) The evolution of the orientation of the rotational axis $\omega$ and the AP axis ${\widehat{\mathbf{\text{d}}}}_3$ is described using the spherical coordinates $(\theta ,\varphi )$ in the inertial frame $E$. The time evolution of the quantities (F-G) is given in Movie 9.

Figure 7:. Proposed biophysical mechanism for egg chamber rotation.

Fat2 operates in a mechanochemical feedback loop, in which cell motion polarizes Fat2 to cells’ trailing edges, which in turn stabilizes and aligns individual cell crawling forces. Rigid-body dynamics of the egg chamber synchronize the individual movements of the follicle cells for collective migration. Tissue geometry provides a mechanical cue that ensures rotation around the AP axis. Synergy between the mechanosensitive behavior of Fat2, egg chamber rigid-body dynamics, and egg chamber geometry initiates and maintains the rotational migration of the follicle cells.