Contact repulsion controls the dispersion and final distribution of Cajal-Retzius cells

Abstract

Cajal-Retzius (CR) cells play a fundamental role in the development of the mammalian cerebral cortex. They control the formation of cortical layers by regulating the migration of pyramidal cells through the release of Reelin. The function of CR cells critically depends on their regular distribution throughout the surface of the cortex, but little is known about the events controlling this phenomenon. Using time-lapse video microscopy in vivo and in vitro, we found that movement of CR cells is regulated by repulsive interactions, which leads to their random dispersion throughout the cortical surface. Mathematical modeling reveals that contact repulsion is both necessary and sufficient for this process, which demonstrates that complex neuronal assemblies may emerge during development through stochastic events. At the molecular level, we found that contact repulsion is mediated by Eph/ephrin interactions. Our observations reveal a mechanism that controls the even distribution of neurons in the developing brain.

Figure 1. Migrating CR cells undergo contact repulsion

(A) Schematic diagram of the distribution of CR cells from different origins in the embryonic cortex at E12.5, adapted from Bielle et al. (2005). (B) Whole-mount telencephalic hemispheres showing the expression of Reelin mRNA. (C) High resolution time-lapse sequence of two migrating CR cells in a flat whole-mount preparation of the E12.5 cortex. The arrow points to a contact between both cells. (D) Low magnification time-lapse sequence of CR cells migrating away from an E12.5 cortical hem explant. (E) Tracks of migrating CR cells and OB interneurons in culture. (F) Directionality indexes for CR cells = 0.3696 ± 0.0184 (n = 93 cells from 6 different experiments) and olfactory bulb interneurons = 0.6681 ± 0.0366 (n = 57 cells from 3 different experiments), Mann-Whitney test: ***p ≤ 0.001. (G and I) High-resolution time-lapse images of the collision between CR cells (pseudocolored) from the same (G) or different origins (I). CR cells derived from the pallial septum (PS) are marked with a gray dot. (H and J) Velocity vectors for CR cells contacts in vitro, with cells from the same (H) or different (J) origin. The black arrow indicates the initial velocity vector. The angle of collision and the velocity vectors were calculated after initial trajectory alignment. Each gray arrow represents one cell. d, dorsal; NCx, neocortex; PS, pallial septum; r, rostral; Th, thalamus; VP, ventral pallium. Scale bars equal 200 μm (B), 25 μm (C), 400 μm (D) and 50 μm (G and I).

Figure 2. Random walking of CR cells during migration

(A) Time-lapse sequence of the random walk behavior of an isolated CR cell in culture. (B) Tracks of isolated CR cells undergoing random walking during migration in culture. (C) Comparison of the directionality index and average turning angle between CR cells undergoing CoRe (gray) and random walking CR cells (blue). Directionality index for CR cells undergoing CoRe (gray) = 0.3080 ± 0.0627 and for random walking CR cells = 0.5296 ± 0.0929 (n = 9). Average turning angle for CR cells undergoing CoRe (gray) = 89.5220 ± 7.9513° and for CR cells undergoing random walking = 65.9046 ± 8.4255° (n = 19 cells from 4 different experiments). t-test, *p ≤ 0.05. (D) Velocity vectors for random walking of CR cells. The black arrow indicates the initial velocity vector. Turning angles and velocity vectors were calculated after initial trajectory alignment. Each arrow represents the value of one cell. Scale bar equals 25 μm.

Figure 3. Modeling CR cell migration

(A and F) Cell distribution profiles obtained through mathematical simulation of cells migrating with (A) or without (F) contact repulsion (CoRe). Cells in both simulations undergo random walking and are present in equal numbers. (B and G) Graphical representation of minimum distances or coverage indexes for the two distributions shown in (A) and (F), respectively. (C to E and H) Normalized quantification of the coverage or average minimum distance between any pixel in the field and the closest cell, mD = 1.8953 ± 0.045 pixels (C), the average difference between the minimum distance between any pixel in the field and the closest cells, std = 2.3616 ± 0.0928 (D), the average minimum distance between any two cells in the field, mDc = 0.5505 ± 0.0088 pixels (E), and the regularity index, RI = 0.8164 ± 0.0169 (H). Values are expressed as relative to measurements in experiments with contact repulsion. n = 25, t-test: ***p ≤ 0.001. (I) Temporal evolution of the coverage or minimum distances among cells migrating with (gray traces) or without (red traces) contact repulsion. Light colors represent individual simulations; dark traces are the average of 25 different experiments. The dotted line at 1000 frames indicates the time represented in (A to H), when the distribution of migrating cells undergoing contact repulsion stabilizes. (J and N) Cell distributions obtained through mathematical simulation of two populations of cells (red and blue) migrating with (J) and without (N) contact repulsion. Cells in both simulations undergo random walking and are present in equal numbers. (K and O) Graphical representation of population overlapping for distributions shown in (J) and (N), respectively. Dark colors indicate areas populated by one of the cell populations; light colors indicate a high degree of overlap between the two cell populations or low cell density. (L and P) Graphical representation of cell overlap between the two cell populations. Light colors represent individual simulations; dark traces are the average of 25 different experiments. (M) Quantification of the segregation index, SI = 0.3897 ± 0.0088 for the two populations of cells in both experimental conditions. Values are expressed as relative to measurements in experiments with contact repulsion. n = 25, t-test: ***p ≤ 0.001. (Q) Temporal evolution of the segregation index between two cell populations migrating with (gray traces) or without (red traces). Light colors represent individual simulations; dark traces are the average of 25 different experiments.

Figure 4. Migrating CR cells express multiple Eph receptors and ephrins

(A) Expression of EphA3, EphA6, EphB1, EphB2, EphB3, EfnA5, EfnB2 and EfnB3 mRNA in individual hem-derived CR cells in culture, identified by the expression of Calretinin. Arrows point to examples of double labeled cells. (B) Distribution of EphA, EphB, ephrinA and ephrinB molecules in the membrane of CR cells in culture revealed by selective binding of EfnA1-, EfnB1-, EphA3- and EphB2-Fc compounds, respectively. Fc fragments alone were used as a negative control of the experiment. CR cells were obtained from the hem of E12.5 ubiquitously expressing GFP embryos. Arrows and arrowheads respectively point to the location of Fc complexes in the leading process and soma of CR cells. CR cells did not stain with control Fc fragments, but consistently stained with each of the Efn- or Eph-Fc fragments (100% stained cells in each case; n = 50 cells per condition). (C) Summary of the expression profile of Eph receptors and ephrins in hem-derived CR cells at E12.5. (+) expressed; (−) not expressed; (?) probably expressed, unconfirmed at the cellular level; (na) not assessed. Scale bars equal 25 μm (A) and 50 μm (B).

Figure 5. Eph/ephrin signaling mediates contact repulsion between CR cells

(A) Schematic of the experimental paradigm. P and D are the proximal and distal sides of the explant, respectively. (B) Dispersion of GFP-expressing hem-derived cells (green) confronted with a second, non-GFP labeled hem explant. CR cells from both explants express Calretinin (red). The aligned arrowheads indicate the intermediate distance between the two explants. The proximal dispersion of migrating cells is sharply reduced in controls, but not when Eph/ephrin signaling is blocked. (C) Quantification of P/D ratio and mean distance migrated by CR cells in the proximal and distal part of the explant. Fc control P/D ratio = 0.6848 ± 0.0659; EphA3/B2-Fc P/D ratio = 1.1292 ± 0.0816, t-test: ***p < 0.001. Mean distance migrated in Fc controls in the proximal part, 563.6833 ± 33.0361 μm; mean distance migrated in EphA3/B2-Fc experiments in the proximal part, 724.3364 ± 35.2412 μm; t-test: **p < 0.01. Mean distance migrated in Fc controls in the distal part, 867.9417 ± 54.8907 μm; mean distance migrated in EphA3/B2-Fc experiments in the distal part, 693.3545 ± 45.9462 μm; t-test: *p < 0.05. We quantified the 30 furthest cells in the proximal and distal regions for 12 control explants and 11 EphA3/B2-Fc treated explants (t-test, *p ≤ 0.05, **p ≤0.01, ***p ≤ 0.001). (D) Quantification of P/D ratios for explants treated with control Fc fragments = 0.6497 ± 0.0953 and EfnA1-Fc = 1.0245 ± 0.0753 (n = 13 explants); control Fc fragments = 0.7826 ± 0.0667 and EfnB1-Fc = 0.9774 ± 0.0568 (n = 19 explants); Fc fragments = 0.7122 ± 0.055 and EphA3-Fc = 0.9236 ± 0.0569 (n = 17 explants); Fc fragments = 0.7662 ± 0.0512 and EphB2-Fc = 0.9920± 0.0790 (n = 17 explants); t-test: *p ≤ 0.05, **p ≤ 0.01. Scale bar equals 400 μm (B)

Figure 6. Dynamic analysis of Eph/ephrin signaling in contact repulsion between CR cells

(A, C and E) High-resolution time-lapse images of the collision between CR cells (pseudocolored) in different experimental conditions. (B, D and F) Tracks of migrating CR cells in culture. (G and I) Directionality indexes (DI), contact times and average after-collision angle for CR cells in different experimental conditions. For Fc treated cells, DI = 0.3723 ± 0.0335; for EphA3+EphB2 Fc treated cells, DI = 0.4907 ± 0.0326, t-test: *p < 0.05. Contact times for Fc treated cells = 2.1428 ± 0.1413 frames; for EphA3+EphB2 Fc treated cells = 3.8333 ± 0.3844 frames, t-test: ***p < 0.001. Average angle after-collision (α); for Fc treated CR cells, α = 72.9765 ± 4.3755°; for EphA3+EphB2 Fc treated cells, α = 54.3540 ± 4.3157°, t-test: **p < 0.01. Control Fc, n = 80 cells; EphA3+EphB2 Fc, n = 80 cells; wild type, n = 93 cells; EphB1/B2/B3 mutant, n = 91 cells; from 4 different experiments in each condition; t-test, *p ≤ 0.05, **p ≤0.01, ***p ≤ 0.001. (H and J) Velocity vectors for CR cells in different experimental conditions. Each arrow represents the average value of 30 cells from an independent experiment. Scale bars equal 25 μm.

Figure 7. Eph/ephrin signaling is required for CR cell cortical coverage in vivo

(A and B) Coronal sections trough the telencephalon of control (A) and EphB1/B2/B3 mutant (B) E12.5 mouse embryos showing the distribution of CR cells by calretinin immunostaining. (A′) and (B′) are high magnification images of the boxed areas in (A) and (B), respectively. Note the irregular distribution of CR cells in mutant embryos (B and B′) compared to controls (A and A′). (C) Representation of the marginal zone coverage by CR cells in coronal sections in both controls (gray) and mutants (red). Note the increase in gaps between CR cells in EphB1/B2/B3 mutants. (D) Percentage area of the cortex occupied by CR cells in E12.5 control (gray) = 8.1558 ± 0.9984% or EphB1/B2/B3 mutant (red) = 5.2788 ± 0.7131% embryos, also divided in rostral (wild type = 14.2609 ± 2.3030%; EphB1/B2/B3 mutant = 9.7243 ± 1.5224%), intermediate (wild type = 4.8366 ± 0.2954%; EphB1/B2/B3 mutant = 3.2513 ± 0.5794%), and caudal (wild type = 5.3700 ± 0.7971%; EphB1/B2/B3 mutant = 2.8605 ± 0.5955%) levels of the cortex (n = 6 animals for each condition). Mann-Whitney test, *p < 0.05, ***p ≤ 0.001. H, hippocampus; LGE, lateral ganglionic eminence; NCx, neocortex; PCx, piriform cortex. Scale bars equal 200 μm.

Figure 8. Eph/ephrin signaling is required for CR cell distribution in vivo

(A and F) Whole-mount telencephalic hemispheres of control (A) and EphB1/B2/B3 mutant (F) E12.5 mouse embryos showing the distribution of CR cells by the expression of Reelin mRNA. (B and G) Binary masks of pictures shown in (A) and (F), respectively. (C and H) Graphical representation of the coverage of the space by CR cells using the minimum distance values obtained from (B) and (G). (D, E and I, J) Graphical representation of the minimum distance or coverage obtained after mathematical simulation of cells migrating with (D and I) or without (E and J) contact repulsion. (K) Quantification of the mean minimum distance (mD) from each pixel of the field to the closest cell and regularity index (RI) in control (gray) and mutant (red) embryos. Control, mD = 5.9000 ± 0.3795 pixels, RI = 1.8379 ± 0.0601; EphB1/B2/B3 mutant, mD = 11.2500 ± 1.6621 pixels, RI = 0.6637 ± 0.2032 (n = 5 wild type and n = 4 mutant embryos, t-test: *p < 0.05, **p ≤0.01). (L) Quantification of coverage or minimum distance and regularity index obtained in silico. Gray bars represent normal cell density with (light gray) or without (dark gray) contact repulsion (CoRe). Red bars represent mutant cell density with (dark red) or without (light red) CoRe. Normal cell density with contact repulsion, mD = 5.62 ± 0.36 pixels, RI = 1.6504 ± 0.0226 (n = 55); normal density without contact repulsion, mD = 9.02 ± 0.84 pixels, RI= 0.62836 ± 0.0324 (n =55); mutant cell density with contact repulsion, mD = 7.92 ± 0.15 pixels, RI = 1.4203 ± 0.0170 (n = 65); mutant cell density without contact repulsion, mD = 10.5 ± 0.98 pixels, RI = 0.8382 ± 0.0186 (n = 65). Two-way ANOVA: ***p ≤ 0.001. NCx, neocortex. Scale bar equals 200 μm.