Figures and data in Migrating mesoderm cells self-organize into a dynamic meshwork structure during chick gastrulation

Figures
Videos
Tables
Additional files

8 figures, 7 videos, 3 tables and 1 additional file

Figures

Figure 1 with 1 supplement

Download asset Open asset

Mesoderm cells move collectively during gastrulation.

(A) Schematic diagram of the chicken embryo at stage HH3. The observation regions are marked by the square boxes in the right panels (A: Anterior, M: Middle, P: Posterior). (B) Experimental procedure. DNA encoding H2B-eGFP was introduced into the cells in the primitive streak at stage HH3 by electroporation. After several hours of incubation, the position of the labeled nuclei was recorded using a multi-photon microscope. (C) Examples of the obtained images of the mesoderm cells expressing H2B-eGFP (upper panels) and reconstructed 3D trajectories (bottom panels). The x, y, and z axes correspond to the mediolateral, anterior-posterior, and dorsoventral axes, respectively. The initial position of the cells is marked by dots on the trajectories. Scale bar: 50 μm. (D) Spatial distribution of progression velocity (arrows) and progression speed (color). (E) Individual cell speed (o) and progression speed (x), and (F) directionality. Each data point of the individual cell speed and the directionality represents the average over the cells and that of the progression speed is the average over the subareas in each region of the six embryos. (G) Mean squared displacement (MSD). Each line corresponds to the MSD in each region of the 6 embryos. (H) Exponent of the MSD plotted in (G). (I) Polar order parameter $φ$ plotted against the radius of the measurement area. The polar order parameter calculated for the cells in the circular areas of a given radius at each time was averaged over the areas and time in each region of the 6 embryos (the crosses and error bars). (J) Mean squared relative distance (MSRD). Each line represents the MSRD in each region of the six embryos. On a time scale larger than about 10 min, the exponent of the MSRD becomes 1. The numbers of cells analyzed are N=1525 (A), 1112 (M), 791(P) (embryo 1), 371 (A), 416 (M), 235 (P) (embryo 2), 398 (A), 388 (M), 316 (P) (embryo 3), 230 (A), 296 (M), 175 (P) (embryo 4), 1386 (A), 1283 (M), 964 (P) (embryo 5), 1040 (A), 1102 (M), 496 (P) (embryo 6).

Figure 1—source data 1 Trajectory data of mesoderm cells used in Figure 1.: https://cdn.elifesciences.org/articles/84749/elife-84749-fig1-data1-v2.xlsx
Download elife-84749-fig1-data1-v2.xlsx

Figure 1—figure supplement 1

Download asset Open asset

Localization and kinematic signatures of mesoderm cells.

(A) Transverse section of an embryo with H2B-eGFP introduced by electroporation. Almost all of the cells expressing H2B-eGFP were located in the mesoderm region between the epiblast and the endoderm. (B) Analysis of the cell motility in the z-direction. Probability density function obtained from the six embryos indicates the frequency of the range of cell motion in z direction, which were obtained as the difference between the maximum and minimum z positions of individual trajectories that were in the image window for more than 60 min. The numbers of cells analyzed are N=1096, 350, 396, 302, 1455, 1044. The thick red line indicates the probability density function averaged over six samples. (C) Individual cell speed and progression speed in the anterior, middle, and posterior regions for the data shown in Figure 1D. The average (x) and the standard deviation (error bars) were shown. (D) Correlation between directionality and the mean squared displacement (MSD) exponent. (E) Auto-correlation function of velocity for each sample.

Figure 2 with 1 supplement

Download asset Open asset

Meshwork structure in mesoderm during gastrulation.

(A) Schematics of the 3D imaging. The white box indicates the imaged area shown in (B). (B) Spatial distribution of the cells in the fixed mesoderm tissue stained for nuclei (cyan) and N-cadherin (green) in the z-projection view (left) and the horizontal section (middle and right). (C) Magnified view of the characteristic meshwork structure in the white box in (B). (D) N-cadherin expression in the middle section of the mesoderm. (E) Binary images of three z-sections in the white box in (D). (F) Persistence diagram (PD) obtained by applying persistent homology analysis to the three z-sections in (E). The pixel size in (E) is 0.192 µm. The points forming a hole branch around the death level ~0 correspond to the holes. Holes are identified as the points with a birth level smaller than –10 μm and a death level larger than –2.5 μm (see Method). The color indicates the multiplicity of the points. (G) Statistics of the radius of holes that appear in the hole branch in the PD and the number of the cells surrounding the holes given that the cell diameter is 10 µm.

Figure 2—source data 1 Numerical data plotted in Figure 2G.: https://cdn.elifesciences.org/articles/84749/elife-84749-fig2-data1-v2.xlsx
Download elife-84749-fig2-data1-v2.xlsx

Figure 2—figure supplement 1

Download asset Open asset

Three-dimensional visualization of holes and their detection by persistent homology.

(A) Transverse section of the embryo. (A1) Horizontal section of the whole chicken embryo at stage HH 4. (A2) Transverse section along the horizontal yellow line in (A1). (A3) Magnified view of the white box in (A2). The holes are marked by the yellow asterisks. Scale bars: (A1) 200 μm; (A2) 50 μm. (B) Correspondence of the points in the persistence diagram (PD) and the holes in the input image obtained by the advanced inverse analysis. The data corresponds to that in Figure 2E upper and 2 F upper. (B1) The points with large persistence in PD form a hole branch around death level ~0. (B2) Magnified view of the red box in B1 showing the hole branch of the points, which correspond to the holes in the input binary image (B3). The numbers assigned to each point in (B2) correspond to those in B3, which confirms the correspondence between the birth-death pairs and the holes. Scale bar, 30 μm in B2.

Figure 3 with 1 supplement

Download asset Open asset

Dynamic meshwork structure.

(A) Successive snapshots obtained from a live image of mesoderm tissue. The position of the six cells at different time points are indicated by the colored asterisk. (B) Persistence diagram (PD) of the three snapshots in (A). The hole branch of the points around death level ~0 away from the diagonal line. The pixel size in (A) is 0.22 µm. Holes are identified as the points with a birth level smaller than –10 μm and a death level larger than –2.5 μm (see Method). The color indicates the multiplicity of the points. (C) The time series of the radius of holes that appear in the hole branch in the PD, and the corresponding number of the cells surrounding the holes that is calculated from the radius under the assumption that the cell diameter is 10 µm. The p-values between any two time points obtained by t-test were larger than 0.05 except for the pairs of 0 min and 8 min, 0 min and 12 min, 0 min and 16 min, 0 min and 20 min, 0 min and 24 min, 4 min and 12 min (p<0.01), and 8 min and 12 min (p<0.05), which might possibly be caused by the small size of the data set. (D) Spatiotemporal diagram of the holes. The holes were dynamic with the appearance (4) and disappearance (2,3), as well as the fusion (5) and fission (2).

Figure 3—source data 1 Numerical data plotted in Figure 3C.: https://cdn.elifesciences.org/articles/84749/elife-84749-fig3-data1-v2.xlsx
Download elife-84749-fig3-data1-v2.xlsx

Figure 3—figure supplement 1

Download asset Open asset

Dynamics of holes.

(A) Image of the time frame t=0 min of the live imaging. The transverse (horizontal) view was obtained at the level indicated by the yellow dotted line in the horizontal (transverse) view. Scale bars, 50 μm. See also Video 4. (B) Change of the meshwork structure over time. The holes in the images of 0 min (red), 15 min (yellow), 25 min (blue), and 35 min (green) are manually traced. (C) Displacement of the contour of the holes traced manually during 35 min. The holes move in the anterior-lateral direction during the observation as indicated by the arrows. Scale bar, 50 μm. (D) Input binarized images used for the persistent homology analysis and the extracted images by using the advanced inverse analysis. The five holes labeled by ①-⑤ are extracted to visualize their time evolution in Figure 3D by stacking them along the t axis.

Figure 4 with 3 supplements

Download asset Open asset

Intercellular adhesion controlling collective mesoderm cell migration.

(A) N-cadherin expression in the mesoderm. N-cadherin was localized at the cell-cell contact sites both in the horizontal section (left) and in the vertical section (right) that surround the holes. Scale bars, 10 μm. (B) Structure of the wild-type N-cadherin and the deletion mutant of N-cadherin consisting of the cytoplasmic domain with myristoylation signal (top). Schematic diagram of the experimental method (bottom). To compare the migration of the mesoderm cells, H2B-eGFP was electroporated on the A side, while the N-cadherin mutant (N-Cad-M) was electroporated on the B side. The N-Cad-M expressing cells were marked by the H2B-mCherry expression. (C) Effects of N-Cad-M overexpression on endogenous cadherin expression. Endogenous N-cadherin (left) and P-cadherin (right) are expressed specifically at the cell-cell contact site in the control mesoderm cells (white arrow heads). In contrast, in the cells expressing N-Cad-M labeled in red, the expression of N-cadherin (left) and P-cadherin (right) were almost disappeared from the cell membrane (yellow arrow heads). Scale bars, 10 μm. (D) N-Cad-M expressing cells tend not to be integrated into the meshwork structure of control mesoderm cells (yellow arrow heads). (D1) and (D2) Magnified images in the white boxes in the top panel. The N-cadherin (N-Cad-M) expressing cells did not participate in the meshwork. Scale bar, 100 μm. (E) Examples of (top) a snapshot of the live imaging and (bottom) trajectories of the mesoderm cells expressing H2B-eGFP (A side) and the N-Cad-M (B side) of the same embryo. The initial position of the cells is marked by dots on the trajectories. (**F–K**) Statistical quantification of the migration behavior of the control and N-Cad-M cells for five embryos. The corresponding statistical quantity of each cell in each embryo is shown in Figure 4—figure supplement 2. The quantities of control and N-Cad-M in the same embryo are linked by the line. (F) Mean of individual cell speed (p=0.47). (G) Mean of directionality (p=0.0095). (H) MSD exponent (p=0.00979). (I) Mean of progression speed (p=0.02). (J) Polar order parameter (p=0.081). (K) Auto-correlation function (ACF) of the direction of collective migration at 10 min (p=0.0075). p-values were obtained using paired t-test of the five embryos. The xy size of the imaged square area of five embryos: 258, 192, 207, 500, 500 µm. The numbers of cells analyzed: N=118 (Control), 119 (Mutant) (embryo 1), 41 (C), 28 (M) (embryo 2), 44 (C), 30 (M) (embryo 3), 253 (C), 290 (M) (embryo 4), 297 (C), 232 (M) (embryo 5).

Figure 4—source data 1 Trajectory data of mesoderm cells used in Figure 4.: https://cdn.elifesciences.org/articles/84749/elife-84749-fig4-data1-v2.xlsx
Download elife-84749-fig4-data1-v2.xlsx

Figure 4—figure supplement 1

Download asset Open asset

Multiphoton images of the control cells and the N-cadherin (N-Cad-M) expressing cells (‘Raw image’ in the left panel).

Cell membrane is marked by GFP-CAAX and extracted by surface function of IMARIS software (“Extracted cells by Surface Function” in the left panel). (B right top panel) A schematic diagram shows the shortest length and the longest length of a cell. a: Length of the shortest principal axis. b: Length of the longest principal axis. The software identifies a cell by considering the object-oriented minimal rectangular box, as shown in red. (B right bottom panel) Aspect ratio of the control cells and the N-Cad-M expressing cells. Error bars are the standard error of mean; N=51 (control), N=37 (N-Cad-M). Asterisk, P<0.001 (t-test).

Figure 4—figure supplement 2

Download asset Open asset

Statistical comparison between control and N-Cadherin mutant expressing cells within individual embryos.

Distribution of (A) individual cell speed, (B) directionality of individual cells, (C) progression speed of 50 µm × 50 µm areas, and (D) polar order parameter of 125 µm × 125 µm, and (E) autocorrelation function (ACF) of the direction of collective migration. (A-D) Each circle (o) plots the temporal average. The average (x) over (**A–B**) the cells and (**C–D**) the areas and their standard error (error bar) were shown. P-values indicated in the graphs were obtained from Wilcoxon rank sum test between the control and mutant cells. In embryo ID 1, 2, and 4, the polar order parameter on the control side was higher than that on the mutant side. In embryo ID 3 and 5, the difference was not statistically significant. (E) The average and the standard error of means of ACF calculated for each subarea (125 µm × 125 µm) are shown.

Figure 4—figure supplement 3

Download asset Open asset

Cell-cell contact behaviors in control and N-Cadherin mutant expressing cells.

(A) Snapshots of the live imaging of the control cells (upper panels) and the N-cadherin (N-Cad-M) expressing cells (lower panels). Cell membrane is marked by GFP-CAAX. The numbers (1–5 in the images of control cells, 1–7 in the images of N-cadherin mutant cells) are assigned to track the cells. Scale bars, 10 μm.

Figure 5 with 1 supplement

Download asset Open asset

Theoretical model of meshwork formation.

(A) Impact of the attractive interaction strength $ϵ_{a t r}$ on the meshwork structure formation. (B) Dependence of the birth level of the holes on the attractive interaction strength $ϵ_{a t r}$ . The error bars indicate the standard error of mean obtained from n=10 independent simulations. (C) Impact of the agent aspect ratio $r$ on the meshwork structure formation. (D) Dependence of the birth level of the holes on the aspect ratio $r$ obtained from the persistent homology analysis. The error bars indicate the standard error of mean obtained from n=10 independent simulations. (E) Alignment of the agents in the aggregates and the meshwork structure. Aspect ratio $r$ = 4. (F) Relation between the nematic direction of agents in the aggregates and the elongation direction of the aggregates. (left) Relationship between the nematic angle and the longitudinal angle of each aggregate. The color of the points represents the aspect ratio of the aggregates. (right) The correlation between the nematic angle $θ_{n}$ and the aggregate longitudinal angle $θ_{a}$ defined by $⟨ \cos 2 (θ_{n} - θ_{a}) ⟩$ as a function of the aggregate density for different aspect ratio $r$ . (G) Impact of the agent supply rate on the meshwork structure formation in the simulation with the agents supplied from the PS boundary on the left. Snapshots (top) and persistence diagrams (PD) (bottom). (H) Impact of the adhesion and the aspect ratio on the meshwork structure formation in the simulation with the agents supplied from the PS boundary. Snapshots (left) and PD (right).

Figure 5—figure supplement 1

Download asset Open asset

Theoretical model for the meshwork formation of chick mesoderm cells.

Schematics of the theoretical model (A) and the potential and force profiles of the attractive interactions (**B,C**).

Figure 6

Download asset Open asset

Changes in the meshwork structure during development.

(A) Spatial distribution of the cells in the mesoderm tissue stained for nuclei (cyan) and N-cadherin (white) at different developmental stages (top) and the corresponding persistence diagrams (PD) (bottom). The persistent homology analysis was performed using binary images. The pixel size is 0.215 μm. (B) Simulation with the supply of the agents where the supply rate increases with time. Snapshots (top) and the corresponding PD. (C) The radius of holes that appear in the hole branch in the PD in (A). (D) The radius of holes that appear in the hole branch in the PD in (B). Holes are identified as the points with a birth level smaller than –5 μm (A) and –10 μm (B) and a death level larger than –2.5 μm (see Method). The color in the PD diagram (AB) indicates the multiplicity of the points.

Figure 6—source data 1 Numerical data plotted in Figure 6C.: https://cdn.elifesciences.org/articles/84749/elife-84749-fig6-data1-v2.xlsx
Download elife-84749-fig6-data1-v2.xlsx
Figure 6—source data 2 Numerical data plotted in Figure 6D.: https://cdn.elifesciences.org/articles/84749/elife-84749-fig6-data2-v2.xlsx
Download elife-84749-fig6-data2-v2.xlsx

Author response image 1

Download asset Open asset

Percolation transition happens as the density increases.

The critical density becomes smaller when the attractive interaction is introduced, but it becomes large again for too high attraction.

Author response image 2

Download asset Open asset

Brachury expression level at different z level of early chick embryo.

Videos

Video 1

Download asset

posterframe for video — Mesoderm cell movements on gastrulating chick embryo.

Left: Nuclei of mesoderm cells are labeled by H2B-eGFP expression (green). Right: Cell trajectories by IMARIS tracking in the anterior, middle and posterior regions. Scale bars, 50 μm.

Video 2

Download asset

Video 3

Download asset

Video 4

Download asset

Video 5

Download asset

Video 6

Download asset

Video 7

Download asset

Tables

Key resources table

Reagent type (species) or resource	Designation	Source or reference	Identifiers	Additional information
Biological sample (Chicken)	Wild Type Chicken Eggs	Shimojima farm, Kanagawa, JP
Biological sample (Chicken)	Tg (pLSi/ΔAeGFP) Chicken Eggs	ABRC, University of Nagoya; Motono et al., 2010
Antibody	anti-E-Cadherin (mouse monoclonal)	BD Transduction Lab	Cat# 610181; RRID:AB_397580	IF(1:1000)
Antibody	Anti-N-cadherin, (rabbit polyclonal)	Takara bio	Cat# M142; RRID:AB_444317	IF(1:300)
Antibody	anti-N-Cadherin/A-41CAM Clone GC-4 (mouse monoclonal)	Sigma-Aldrich	Cat# C2542; RRID:AB_258801	IF(1:50)
Antibody	anti-Mouse IgG, (H+L) Highly Cross-adsorbed Antibody, Alexa Fluor 488 (Goat)	Thermo Fisher	Cat# A-11029; RRID:AB_138404	IF(1:300)
Antibody	anti-Rabbit IgG, (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 488 (Goat)	Thermo Fisher	Cat# A-11034; RRID:AB_2576217	IF(1:300)
Chemical compound, drug	Cellstain DAPI Solution	Dojindo Laboratories	Cat# D523	1:100 dilution
Recombinant DNA reagent	pGEM-T Easy	Promega	Cat# A1360
Recombinant DNA reagent	pCAG-H2B-eGFP	Dr. Hadjantonakis (MSKCC, NY).
Recombinant DNA reagent	pCAG-N-Cad-M-2A-H2B-mCherry	This study		See Materials and methods ‘Generation of chick-N-Cadherin mutant’
Recombinant DNA reagent	pCAG-N-Cad-M-2A- eGFP-CAAX-2A-H2B-mCherry	This study		See Materials and methods ‘Generation of chick-N-Cadherin mutant’
Sequence-based reagent	N-cadherin cloning primer Fw 5’-ATGTGCCGGATAGCGGGAAC-3’	Hokkaido System Science Co., Ltd
Sequence-based reagent	N-cadherin cloning primer Rev 5’- TCAGTCATCACCTCCACCG-3’	Hokkaido System Science Co., Ltd
Sequence-based reagent	N-Cad-M primer 1 Fw 5’- ATGGGTTCTTCTAAATCTAAACCAAAAGATCCATCTCAACGTATGAAGCGCCGTGATAAGG-3’	Fasmac
Sequence-based reagent	N-Cad-M primer 1 Rev 5’- GTCATCACCTCCACCGTAC-3’	Fasmac
Sequence-based reagent	N-Cad-M primer 2 Fw 5’- GCGGCCGCGGATCCGCATGCGCCACCATGGGTTCTTCT-3’	Thermo Fisher
Sequence-based reagent	N-Cad-M primer 2 Rev 5’- TTGCTCACCATAACGCATGCTTTAGGTCCAGGGTTCTCC-3’	Thermo Fisher
Software, algorithm	HomCloud	Obayashi et al., 2018	Ver. 3.0.1
Software, algorithm	IMARIS	Oxford instruments, UK	Ver 9.5.1
Software, algorithm	Matlab	Mathworks Inc, Natick, MA	R2024b

Table 1

The parameters used in the numerical simulation.

Variables	Symbols	Values
Cell diameter	$d$	1
Cell head and tail length	$l_{c}$	$d (r - 1)$
Cell stretching elasticity	$κ^{s t r}$	10
Cell-cell attraction strength	$ϵ^{a t r}$	0.001
Cell-cell repulsion strength	$ϵ^{r e p}$	0.1
Cell-cell interaction length	$σ^{c c}$	$d$
Width of cell-cell attraction well	$ξ^{c c}$	$d / 2$
Effective viscosity	$η$	0.1
Temperature	$k_{B} T$	0.004142
Noise strength	$σ$	$2 k_{B} T γ$
Number of cells	$N_{c e l l}$	1600

Table 2

The parameters used in the numerical simulation with the agent supply.

Variables	Symbols	Values
Self-propulsion force	$f^{a c t}$	0.01
Repulsive strength from source	$k^{s o u r c e}$	0.1
Chemotactic force	$f^{c h e m o t a x i s}$	0.004
Simulation box size in x	$L_{x}$	60
Simulation box size in y	$L_{y}$	40