In vivo confinement promotes collective migration of neural crest cells

Abstract

Collective cell migration is fundamental throughout development and in many diseases. Spatial confinement using micropatterns has been shown to promote collective cell migration in vitro, but its effect in vivo remains unclear. Combining computational and experimental approaches, we show that the in vivo collective migration of neural crest cells (NCCs) depends on such confinement. We demonstrate that confinement may be imposed by the spatiotemporal distribution of a nonpermissive substrate provided by versican, an extracellular matrix molecule previously proposed to have contrasting roles: barrier or promoter of NCC migration. We resolve the controversy by demonstrating that versican works as an inhibitor of NCC migration and also acts as a guiding cue by forming exclusionary boundaries. Our model predicts an optimal number of cells in a given confinement width to allow for directional migration. This optimum coincides with the width of neural crest migratory streams analyzed across different species, proposing an explanation for the highly conserved nature of NCC streams during development.

Figure 1.

Versican V0–1 forms a delimiting boundary around migrating NCCs. (A) qPCR of versican isoforms. Bars, mean; error, SEM. (B) Western blot of V0–1 and V3 isoforms. (C–G) Embryos showing expression of versican (C) and the NC marker Twist (D) using whole-mount in situ hybridization (ISH), and corresponding sections (E and F), with a scheme summarizing the sections (G). Sections on E and F have been enhanced by blurring and squaring the separated ISH signals and overlaid in pseudo-color on the section background. (H–L) NCC explants cultured on fibronectin (Fn) + BSA or Fn + versican barriers invade the barrier area only in absence of versican. (J and K) In vitro cluster trajectories with barrier boundary marked as solid line; n = 10; color: time. (L) Percentage of cells invading the barrier (n = 3 independent experiments with n = 251 and n = 336 explants in total for BSA and versican, respectively). Bars, mean; error, SEM. (M–Q) ISH for Twist of embryos grafted with PBS (M) or versican-soaked beads (N), with the corresponding sections (O and P), and inhibition of NC migration (Q; mean inhibition, PBS: n = 5/29, versican: n = 25/32 embryos). Arrowheads show migrating NCC and asterisks indicate grafted bead. **, P < 0.01.

Figure 2.

Versican is required for normal NC migration. (A and B) qPCR (A) and Western blot (B) analysis of CoMO- and VsMO-injected embryos. (C–E) ISH of Twist in embryos injected with either CoMO (C) or VsMO (D), and inhibition of migration (E; C: n = 91, D: n = 135). (F and G) Graft experiments: VsMO + fluorescein-dextran (FDx)–injected NCC grafted into control host (F; 76% of migration, n = 17) or CoMO + FDx NCC grafted into VsMO-injected host (G; 13% of migration, n = 15). (H and I) ISH against Krox20, showing an uninjected versus VsMO-injected side of a stage 24 embryo (H, dorsal view; n = 53), and number of ectopic NCC (I; n = 51). Bars, mean; error, SEM. Arrowhead represents a migrating NC. **, P < 0.01.

Figure 3.

Computational model of NC migration. (A) Schematic embryo during NCC migration in the Xenopus head. (B–D) Cell interactions in the CPM: CIL (B), CoA (C), and dorsal and lateral confinements (D). (E) CPM configurations for versions of NC migration in constrained geometries with and without CIL and CoA. (F and J) Migration efficiencies. Error bars: min-max values, boxes: quartiles; central value: median; n = 50 simulations. Significance compared with relevant control, where significance bars (purple) compare data from the same conditions. (G) CPM configuration in unconstrained geometry. (H) Rules of the DEM model of NCC migration. (I) Configuration of the DEM model with and without confinement. ***, P < 0.001; ns, not significant.

Figure 4.

Confinement by versican enhances collective NC migration. (A–D) Lateral view of stage 27 embryos of wild-type NC (nuclear-GFP in cyan) transplanted into a CoMO-injected (A, n = 21) or VsMO-injected host (C, n = 24; Video 3), with cell trajectories (B and D). (E–H) In silico cell migration with and without confinement with cell trajectories (n = 50). (I–K) Comparison of cell speed, persistence, and direction of migration in vivo and in silico. (L–O) NC cluster migration in vitro. Frames of time-lapse movie of control NC cluster in versican confinement (L, green: versican, red: NC nuclei; n = 20) or without confinement (N; Video 2, n = 25) and cluster trajectories (M and O). (P–S) Simulations imitating the in vitro geometries (n = 50). (T–V) Comparison of cluster speed, persistence, and direction of migration in vitro and in silico. Error bars, min–max, boxes: quartiles; bar, median. ***, P < 0.001; ns, not significant.

Figure 5.

Effect of confinement width and cluster size. (A) Schematic illustration and examples of maximal projection of a Z-confocal stack used to measure the number of cells in NC streams of different widths. White line illustrates how the stream width was measured. Width (W) and number of cells (N) are shown at the bottom, with SD in parentheses; ne, number of embryos analyzed. (B) Snapshots of simulations with increasing confinement widths (w, cell diameters) and cluster size (N, number of cells). (C and D) Cluster persistences and transport ratio (percentage of cells in the cluster migrating at least 150 µm away from the cluster’s edge in 5 h). Heatmaps represent median values from n = 50 simulations. Dots show experimental values for NCs migrating in vivo in zebrafish and Xenopus embryos. Numbers on dots correspond to the same numbers shown in A.